Teses E Dissertações Eletrônicas Na Biblioteca Digital Da Ufg

UNIVERSIDADE FEDERAL DE GOIÁS E UNIVERSIDADE DE ALCALÁ Programa de Pós-Graduação em Ecologia e Evolução e Doctorado en Ecología. Conservación y Restauración de Ecosistemas

DO SURGIMENTO À DESCRIÇÃO:

COMO O TEMPO E A EXTINÇÃO ESTÃO RELACIONADOS

BRUNO VILELA DE MORAES E SILVA

Goiás, 2016

TERMO DE CIÊNCIA E DE AUTORIZAÇÃO PARA DISPONIBILIZAR AS TESES E DISSERTAÇÕES ELETRÔNICAS NA BIBLIOTECA DIGITAL DA UFG

Na qualidade de titular dos direitos de autor, autorizo a Universidade Federal de Goiás (UFG) a disponibilizar, gratuitamente, por meio da Biblioteca Digital de Teses e Dissertações (BDTD/UFG), regulamentada pela Resolução CEPEC nº 832/2007, sem ressarcimento dos direitos autorais, de acordo com a Lei nº 9610/98, o documento conforme permissões assinaladas abaixo, para fins de leitura, impressão e/ou download, a título de divulgação da produção científica brasileira, a partir desta data.

1. Identificação do material bibliográfico: [ ] Dissertação [ X] Tese

2. Identificação da Tese ou Dissertação

Nome completo do autor: Bruno Vilela de Moraes e Silva

Título do trabalho: Do surgimento à descrição: como o tempo e a extinção estão relacionados

3. Informações de acesso ao documento:

Concorda com a liberação total do documento [ X ] SIM [ ] NÃO1

Havendo concordância com a disponibilização eletrônica, torna-se imprescindível o envio do(s) arquivo(s) em formato digital PDF da tese ou dissertação.

______Data: 31 / 01 / 2017 Assinatura do (a) autor (a)

Ficha catalográfica elaborada automaticamente com os dados fornecidos pelo(a) autor(a), sob orientação do Sibi/UFG.

Vilela de Moraes e Silva, Bruno Do surgimento à descrição: como o tempo e a extinção estão relacionados [manuscrito] / Bruno Vilela de Moraes e Silva. - 2016. X, 129 f.: il.

Orientador: Profa. Dra. Levi Carina Terribile; co-orientador Dr. Miguel Ángel Rodriguez. Tese (Doutorado) - Universidade Federal de Goiás, Instituto de Ciências Biológicas (ICB) , Programa de Pós-Graduação em Ecologia e Evolução, Cidade de Goiás, 2016. Bibliografia. Anexos. Apêndice. Inclui mapas, gráfico, tabelas.

1. Ano de descrição . 2. Dados insuficientes. 3. Espécies ameaçadas. 4. Idade do clado. 5. Listas vermelhas. I. Carina Terribile, Levi, orient. II. Ángel Rodriguez, Miguel, co-orient. III. Título.

Universidade Federal de Goiás Universidad de Alcalá Instituto de Ciências Biológicas Departamento de Ciencias de la Vida Programa de Pós-Graduação em Ecologia Doctorado en Ecología. Conservación y e Evolução Restauración de Ecosistemas

BRUNO VILELA DE MORAES E SILVA

DO SURGIMENTO À DESCRIÇÃO:

COMO O TEMPO E A EXTINÇÃO ESTÃO RELACIONADOS

Tese apresentada ao Programa de Pós-graduação stricto sensu em Ecologia & Evolução/ICB/UFG e Doctorado em Ecolgía: Conservación y Restauracón de Ecosistemas/UAH como parte dos requisitos para obtenção do título de doutor por ambas universidades, dado acordo de cotutela.

Orientadora(UFG): Dra. Levi Carina Terribile

Orientador(UAH): Dr. Miguel Ángel Rodriguez

Goiânia-GO, março de 2016

III

Dedico essa tese à Amélie Ahimsa, que me manteve acordado tempo suficiente para que eu pudesse conclui-la, e à sua mãe, que carinhosamente fazia ela dormir novamente para que eu pudesse trabalhar...

“O principezinho estava agora pálido de cólera.

- Há milhões e milhões de anos que as flores fabricam espinhos. Há milhões e milhões de anos que os carneiros as comem, apesar de tudo. E não será sério procurar compreender por que perdem tanto tempo fabricando espinhos inúteis? Não terá importância a guerra dos carneiros e das flores? Não será mais importante que as contas do tal sujeito? E se eu, por minha vez, conheço uma flor única no mundo, que só existe no meu planeta, e que um belo dia um carneirinho pode liquidar num só golpe, sem avaliar o que faz, - isto não tem importância?!”

Antoine de Saint-Exupéry

AGRADECIMENTOS

Longe de ser uma jornada solitária, ou uma monografia por assim dizer, o processo de pensar, desenvolver e escrever essa tese envolveu diversas pessoas, com as quais tenho a obrigação e a alegria de compartilhar os frutos desse trabalho.

Nada mais justo de que iniciar agradecendo minha orientadora Professora Levi Carina Terribile, que sem ao menos me conhecer, deu seu voto de confiança aceitando me orientar durante esses quatro anos. Sob sua supervisão tive liberdade e ao mesmo tempo suporte para que eu pudesse seguir minhas próprias ideias, testando e explorando as mais diversas possibilidades. Em grande parte, como resultado dessa autonomia que me foi dada, me sinto bastante seguro a seguir meu próprio caminho após o final do doutorado. Sempre pronta para ajudar no que fosse necessário, Carina se mostrou uma excelente orientadora e sem ela essa tese não haveria nem ao menos iniciado.

A seguinte pessoa a quem agradeço é ao Professor Miguel Ángel Rodriguez, o qual me deu a oportunidade de realizar essa tese em parceria com a Universidade de Alcalá. Miguel me contagiou com seu entusiasmo pelas questões centrais da biogeografia. Ao mesmo tempo que me permitiu trabalhar em minhas próprias perguntas, também me incentivou a colaborar nas mais diversas pesquisas desenvolvidas em seu laboratório. Sempre animado, não deixava passar uma oportunidade de discutir novas ideias, mas nunca se colocava em condição superior em relação a seus alunos, aceitando nossas opiniões e nos aconselhando quando necessário. Só tenho a agradecer seus ensinamentos e o bom tempo que passei na Espanha.

Sempre me considerei uma pessoa de sorte. Isso se confirmou ao chegar em Goiânia e acontecer de dividir apartamento com uma pessoa que viria a ser um grande amigo, o qual considero também o terceiro orientador dessa tese, o Dr. Fabricio Villalobos. Das inúmeras e interessantes discussões que tivemos, surgiram algumas das ideias centrais dessa tese. Fabrício foi um personagem essencial desde a elaboração até a publicação de três dos capítulos aqui apresentados. Com sua personalidade persistente, grande vontade de trabalhar

e colaborar, nos divertimos explorando perguntas, dados, análises e resultados. Com ele aprendi muito do que sei, agradeço seu apoio e amizade durante esses quatro anos.

Antes de mais nada, é preciso agradecer as duas outras pessoas que tiveram coautoria em um dos capítulos dessa tese, o Professor Juan Carlos Moreno Saiz e o Dr. Rafael Molina‐Venegas. Ambos participaram ativamente do desenvolvimento do terceiro capítulo dessa tese compartilhando seu conhecimento sobre a história, taxonomia e relações filogenéticas das plantas da região ibérica.

Durante o tempo que passei em Goiânia, tive o privilégio de poder trabalhar no Laboratório de Ecologia Teórica e Síntese (LETS). Agradeço em especial aos professores José Alexandre Diniz-Filho, Luis Maurício Bini, Thiago Rangel e Paulo de Marco Júnior por essa oportunidade. No LETS foi realizado grande parte do trabalho apresentado aqui, o qual teve influência de vários dos meus colegas de laboratório. É pensando nessas pessoas que fizeram e fazem parte desse laboratório, que nomeamos (junto com Fabrício) o pacote de R aqui apresentado no quarto capítulo, letsR.

Agradeço ao Professor Adriano Sanchez Melo pelos ensinamentos de R e bons conselhos ao longo desses anos. O aprendizado dessa linguagem de programação foi a base para todas as análises dessa tese. Em seu nome agradeço a todos os professores com quem tive oportunidade de aprender durante esses quatro anos. Aproveito para também agradecer a minha amiga, Dra. Sara Varela, que me forneceu o conhecimento necessário para transformar tais análises em um pacote de R (e ao Javi pelas aulas de Github!).

Devo reconhecer que esclarecimento de conceitos, discussão de ideias e diversas sugestões nos trabalhos vieram de muitos colegas da UFG, sejam eles no antigo banquinho do ócio, nos cafezinhos da vida, no velho sofá do LETS, no famoso Gurupi, no eterno Barbosa Lima ou durante os tradicionais churrascos. Assim, agradeço aos meu amigos e colegas que tornaram esses anos de doutorado tão interessantes: Sidney Gouveia, Ricardo Dobrovolski, Diogo Provete, Karen Neves, Lucas Jardim/Doug boy, Luciano Sgarbi, Marcos Vieira, Jesus Pinto Ledezma (que corrigiu meu resumo em espanhol!!), Julio Grandez, Bruno Barreto, Guilherme de Oliveira, Lorena Mendes, Daniel Paiva Silva, Vanessa Lopes, Ludmila Rattis, Núbia Marques, Clarissa Ruas, Andreia Isaac, Paola

VII

Nobre, Fabio Carvalho, Thiago Bernardi, Leandro Maracahipes, Karina Dias, Tulio Max, Carol Caiado, Fernanda Martins, Diogo Samia, Frederico Faleiro, Denis Nogueira, Davi Mello, Matheus Ribeiro, Talita Braga... e muitos outros colegas do ppg EcoEvol que não estão citados diretamente aqui por conta da minha péssima memória.

Agradeço a todos os “ecofísicos” de Alcalá pelos bons momentos na Espanha! Em especial aos meus companheiros de laboratório e vizinhos de lab Andrei Toca, Laura Fernandez, Isabel Rivas, Silvia Medina, Luciano Pataro, Mila Ferreira, Joaquin Calatayud y Andrea Briega pelas discussões científicas e não-científicas. Além destes, agradeço aos professores da UAH e sua equipe técnica que cuidaram de tudo para que a tese fosse desenvolvida em cotutela com esta universidade.

Agradeço também as pessoas que não estiveram diretamente envolvidas nesse trabalho, mas que com seu apoio e suporte emocional me permitiram realizá-lo da melhor maneira possível. Incluo aí toda minha família, em especial aos meu pais, Ronaldo e Sandra, a todos os meus amigos e as minhas queridas Cacá, Dinha e Jó. Também não posso deixar de agradecer aos grandes amigos do “Piso 6” e filial, pela convivência na Espanha! Assim como ao Vidal e Artur por nos acolher em Goiânia no último ano. Não menos importante, agradeço às minhas γλυκουλες, Effie (que também corrigiu o inglês de alguns dos trabalhos) e Amélie pela inspiração proporcionada.

Por fim, agradeço a CAPES pelo suporte financeiro por meio de uma bolsa de doutorado e doutorado sanduíche que me permitiram uma dedicação exclusiva ao doutorado.

Escrever os agradecimentos para quatro anos de trabalho é uma tarefa muito difícil, assim, peço desculpas aos que ficaram de fora dessa lista, mas espero que todas as pessoas citadas aqui ou não, sintam-se parte dessa tese e também de minha formação ao longo desse doutorado. A vocês, meu muito obrigado!!!

VIII

SUMÁRIO

RESUMO...... 10 RESUMEN ...... 11

INTRODUÇÃO GERAL: Uma visão panorâmica do nosso conhecimento sobre os fatores que levam uma espécie a extinção ...... 12 ESTABELECENDO E CLASSIFICANDO AS CAUSAS DA EXTINÇÃO ...... 13 ADICIONANDO COMPLEXIDADE: INTERAÇÕES ...... 17 AVALIANDO AS CAUSAS NA PRÁTICA: O PROBLEMA DO VIÉS DE INFORMAÇÃO ...... 19 APRESENTAÇÃO DA TESE...... 22 REFERÊNCIAS ...... 24

CAPÍTULO 1: Body size, extinction risk and knowledge bias in New World snakes ...... 32 ABSTRACT ...... 33 INTRODUCTION ...... 34 MATERIAL AND METHODS ...... 36 The data...... 36 Numerical methods ...... 37 RESULTS ...... 40 Snake richness, extinction risks and geographic patterns ...... 40 Risk categories and species’ description years ...... 40 Snake body size and extinction risk categories ...... 41 DISCUSSION ...... 42 ACKNOWLEDGEMENTS ...... 45 REFERENCES ...... 46 FIGURES ...... 52

CAPÍTULO 2: The loss of the unknown: or how higher extinction risks are biased towards newer species to science ...... 56 ABSTRACT ...... 57 INTRODUCTION ...... 58 METHODS ...... 60 Species data...... 60 Extinction risk vs. Description year ...... 61 Effect of geographical range and population size ...... 61 RESULTS ...... 62 DISCUSSION ...... 63 ACKNOWLEDGEMENTS ...... 65 REFERENCES ...... 66 TABLES ...... 69 FIGURES ...... 71 SUPPORTING INFORMATION ...... 73

CAPÍTULO 3: Biogeography underlies the effects of evolutionary history on current extinction risks in the Iberian flora ...... 74 ABSTRACT ...... 75 INTRODUCTION ...... 76 METHODS ...... 80 Species list and extinction risk data ...... 80 Phylogeny ...... 80 Phylogenetic signal ...... 81 Phylogenetic traits ...... 83 Statistical analysis...... 84 Phylogenetic diversity loss scenario ...... 85 Data availability and analysis reproducibility ...... 86 RESULTS ...... 86 Phylogenetic signal in extinction risk ...... 86 Phylogenetic traits effect on the extinction risk ...... 87 Disproportional loss of evolutionary history ...... 88 DISCUSSION ...... 88 Phylogenetic clustering in extinction risk in Eastern Andalusia but not in the whole Spanish Iberian Peninsula ...... 88 Biogeography intermediates the phylogenetic trait effects on the extinction risk ...... 89 The expected impact on evolutionary history ...... 92 CONCLUSIONS ...... 92 ACKNOWLEDGEMENTS ...... 93 REFERENCES ...... 93 TABLES ...... 100 FIGURES ...... 103 SUPPORTING INFORMATION ...... 108

CAPÍTULO 4: letsR: a new R package for data handling and analysis in macroecology 109 SUMMARY ...... 110 INTRODUCTION ...... 111 LETSR PACKAGE ...... 112 PresenceAbsence class ...... 113 Memory and time consuption ...... 114 EXAMPLE: SPATIAL PATTERN OF DESCRIPTION YEAR IN TAILLESS AMPHIBIANS (AMPHIBIA: ANURA) ...... 114 CONCLUSIONS ...... 117 ACKNOWLEDGEMENTS ...... 117 DATA ACCESSIBILITY ...... 118 REFERENCES ...... 119 TABLES ...... 123 FIGURES ...... 124

CONCLUSÃO GERAL ...... 125

RESUMO

A extinção é um processo chave na evolução das espécies e, como a maior parte dos processos biológicos, envolve uma complexidade causada pela multiplicidade de fatores envolvidos e suas correspondentes interações. Nessa tese procuramos estudar nos dois primeiros capítulos parte dessas causas, particularmente como o tempo compreendido entre a descrição científica de uma espécie até a atualidade pode influenciar o nosso conhecimento sobre a mesma e seu risco de extinção associado. Partindo das serpentes do Novo Mundo (primeiro capítulo) e depois ampliando os resultados a diversos grupos animais em um nível global (segundo capítulo), focamos em como o tamanho do corpo, a distribuição geográfica e a abundância estão gerando uma correlação entre o ano em que uma espécie foi descrita, sua informação associada e seu risco de extinção. Mostramos como isso pode ocasionar potenciais erros em estudos correlativos e prováveis consequências negativas para o conhecimento científico da biodiversidade, já que os resultados indicam que estamos perdendo as espécies menos conhecidas para a ciência. No terceiro capítulo, mostramos como a idade de uma espécie, juntamente com outras características evolutivas, influenciam o risco de extinção atual em plantas ibéricas e mediterrâneas, e como essa relação pode levar a uma perda desproporcional de história evolutiva na região. Apresentamos uma proposta teórica de como o padrão filogenético do risco de extinção observado pode ser explicado pela interação com a história biogeográfica. Por fim, agregamos uma série de ferramentas de obtenção, manipulação e análise dados macroecológicos desenvolvidas ao longo dos capítulos dessa tese em um pacote de R, apresentado no último capítulo. Exemplificamos a sua funcionalidade ao revelar o padrão global do ano de descrição em anfíbios, mostrando o papel relativo da influência humana e do tamanho da distribuição geográfica.

Palavras-chave: Ano de descrição; Dados Insuficientes; Espécies ameaçadas; Idade do clado; Listas vermelhas; Viés de informação.

RESUMEN

La extinción es un proceso central en la evolución de las especies y como todo proceso biológico su complejidad es causada por la conjunción de factores actuantes y sus correspondientes interacciones. En esta tesis buscamos estudiar parte de esas causas, particularmente como el tiempo comprendido entre la descripción científica de una especie hasta la actualidad puede influir en nuestro conocimiento sobre la misma y su riesgo de extinción asociado. Desde las serpientes del nuevo mundo (primer capítulo) hasta diversos grupos animales de todo el mundo (segundo capítulo), enfocamos en como el tamaño del cuerpo, el rango de distribución geográfica y la abundancia relativa están generando una correlación entre el año en que una especie fue descripta, la cantidad de información y el riesgo de extinción. Mostramos como esas correlaciones pueden ocasionar errores en estudios correlativos y probables consecuencias negativas para el conocimiento de la biodiversidad, puesto que los resultados demuestran que se están perdiendo especies menos conocidas para la ciencia. En el tercer capítulo, indicamos como la edad del clado, juntamente con otras características evolutivas, influencian el riesgo de extinción actual en plantas ibéricas y mediterráneas, y como este efecto pude llevar a una pérdida desproporcional de historia evolutiva en la región. Presentamos un modelo teórico de como los patrones evolutivos del riesgo de extinción observados pueden ser explicados por la interacción con la historia biogeográfica. Finalmente, agregamos herramientas para la obtención, manipulación y análisis de datos, que fueron usados en los diferentes capítulos de esta tesis, en un paquete de R, presentado en el último capítulo. Se ejemplifican sus funcionalidades al estudiar el patrón global del año de descripción en anfibios, mostrando el papel relativo de la influencia humana y el rango de distribución.

Palabras-clave: Año de descripción; Datos Insuficientes; Edad del clado; Especies amenazadas; Listas rojas; Sesgos de información.

INTRODUÇÃO GERAL: UMA VISÃO PANORÂMICA DO NOSSO CONHECIMENTO

SOBRE OS FATORES QUE LEVAM UMA ESPÉCIE A EXTINÇÃO

‘Extinção’, atualmente um conceito biológico amplamente conhecido e aceito pela maioria das pessoas como o desaparecimento completo dos indivíduos de uma espécie. Porém, nem sempre isso foi assim. Até o final do século XVIII, a ideia da extinção de uma espécie seria considerada um absurdo, ao menos pelo público em geral. O trabalho considerado por muitos como o primeiro a confirmar a extinção de uma espécie foi realizado pelo então auxiliar do curador da coleção do Museu de História Natural de Paris, George Cuvier [1]. Ao comparar a mandíbula de um fóssil do norte da Ásia ao de elefantes viventes, ele percebeu que se tratava de uma nova espécie, uma jamais vista nos tempos modernos. Como seria muito difícil que um animal tão grande estivesse escondido por tanto tempo, em 1795, a extinção do mamute era considerada como um fato [2].

Não demorou muito para que surgissem os primeiros debates acerca do que havia acontecido para que essa e outras espécies tivessem desaparecido. Seguindo o paradigma criacionista estabelecido na época, a primeira causa considerada foi o dilúvio bíblico. Aparentemente, algumas espécies haviam chegado atrasadas ou foram impedidas de entrar na arca pelo próprio Noé [3]. Contudo, a posição dos fósseis em diferentes porções dos estratos geológicos (já conhecidos na época, inclusive por Cuvier) indicavam que as extinções ocorreram em diferentes períodos, e, portanto, um único evento não era capaz de explicar todas as extinções [1]. Havia então uma necessidade de se buscar novas causas, dessa vez que não fossem estabelecidas pela fé, mas sim por fatos científicos. Podemos situar esse ponto da história como início da busca por uma resposta à pergunta que engloba essa própria tese: o que leva uma espécie a ser extinta?

Mais de 200 anos de avanços teóricos e tecnológicos foram necessários até poder estabelecer o que realmente aconteceu ao mamute. Aparentemente, as mudanças climáticas do fim do pleistoceno restringiram a espécie a atual Sibéria, lá, a caça intensiva pelas populações humanas foram o golpe final para eliminar as últimas populações desses animais

[4]. A ironia é que depois de toda a discussão a resposta era simples, nós fomos os responsáveis finais pelo desaparecimento do mamute.

Certamente os mamutes não são os únicos incluídos em nossa triste lista, que conta com ao menos 322 espécies de vertebrados terrestres extintos pela humanidade [5]. O domínio da agricultura pelo Homo sapiens permitiu o estabelecimento das primeiras cidades, desencadeando uma rápida evolução cultural que culminaria na revolução industrial no século XIX. É após esse período, que grande parte das principais transformações do meio ambiente ocorre, inclusive da própria condição climática do planeta [6, 7]. Essas modificações do ambiente são apontadas como a principal causa das extinções de espécies nos dias de hoje [5, 8]. Estimativas comparando as taxas normais de extinção em fósseis às taxas apresentadas atualmente indicam que podemos estar presenciando um novo evento de extinção em massa [9]. Durante os ~3,5 bilhões de história da vida, apenas em cinco momentos anteriores o mundo passou por tamanha perda de biodiversidade ocasionadas por desastres naturais de proporções globais, e ao que tudo indica, a sexta será causada pela humanidade [10] (o que nos coloca na categoria de desastres!).

A extinção de espécies deve acarretar uma série de problemas para a nossa própria sociedade. É esperado que muitos dos serviços providos pelos ecossistemas, como polinização, controle de pragas e doenças, conservação de nascentes, contenção de inundações, manutenção do clima, entre incontáveis outros serviços sejam negativamente afetados pela perda de biodiversidade [11, 12]. Isso terá consequências devastadoras para a economia, saúde e cultura do mundo inteiro, como já é possível verificar em muitos locais [13]. Muito além da questão prática, conservar a biodiversidade da melhor maneira possível é uma questão de ética para com a vida. A tão falada crise da biodiversidade precisa ser combatida, mas antes de mais nada precisa ser entendida, e o processo central dessa crise são as extinções e suas causas.

Estabelecendo e classificando as causas da extinção

Apesar da alarmante influência humana na atual taxa de desaparecimento de espécies, é preciso recordar que a extinção se trata de um processo natural. Tão comum, que se estima que 99% das espécies que uma vez habitaram esse planeta foram extintas [14]. Afinal, o

conceito de extinção é parte fundamental no mecanismo de evolução por seleção natural, introduzido por Darwin e Wallace no século XIX [15]. Na verdade, pode-se estabelecer a seleção natural como a única explicação para a extinção. Qualquer intuito de identificar uma causa para os eventos de extinções (no passado, presente ou futuro) deve estar munido da ideia de adaptação as condições abióticas e bióticas do ambiente ao qual a espécie estava/está/estará submetida [16].

Comumente quando se fala em causas de extinção, pensamos principalmente nos fatores abióticos. Isso provavelmente vem da ideia de que extinções são ocorrências atípicas e estão ligadas a alterações do ambiente como consequência de atividades humanas, por exemplo. Como veremos aqui, os fatores abióticos são apenas um grupo de variáveis entre outras que podem estar determinando a extinção de uma espécie. Estes fatores são incluídos em uma categoria frequentemente chamada de fatores extrínsecos [14, 17-19]. Assim chamados por não estar associados a espécie per si, mas ao ambiente onde ela se encontra, compreendendo qualquer variável que pode ser medida em um local ou região, como a altitude, tipo de solo, temperatura, características da paisagem, etc.

Na categoria de fatores extrínsecos, pode-se incluir também as relações bióticas, isto é, causas de extinção que estão relacionadas as interações entre as espécies. Note que incluí esses fatores como extrínsecos porque essas relações em última instância dependem do pool de espécies do local, mas não está claro até que ponto essa não é uma variável da própria espécie, como por exemplo em análise de redes (uma boa discussão seria estabelecer essa dualidade das variáveis bióticas). Eventos de extinção causados por interações bióticas podem incluir interações com parasitas ou outras espécies causadoras de doenças [e.g. 20], exclusão competitiva/predatória [e.g. 15, 21] e extinções em cascata (ou co-extinções) [e.g. 22], todas muitas vezes acionadas por espécies invasoras [23], mudança do ambiente [24] ou por um processo evolutivo natural. Voltando ao caso do mamute, poderíamos dizer que as causas extrínsecas de sua extinção foram as mudanças climáticas (abiótico) e a intensidade da caça humana (nesse caso classificado como biótico, como parte de uma exclusão predatória).

Um tipo especial de fator extrínseco é o espaço em si. O conceito de espaço nesse caso é definido como a área acessível a espécie, o que para algumas será os limites 15

geométricos do continente, para outras do oceano, para outras o próprio planeta, dependendo das limitações fisiológicas e morfológicas da própria espécie [25]. O espaço é importante porque ele determina tanto o pool de fatores extrínsecos aos quais a espécie está submetida [26] e também atua como um limite físico da distribuição geográfica e indiretamente do tamanho da população [27].

As causas extrínsecas de extinção não são totalmente independentes em relação as peculiaridades de cada espécie. Caso contrário, tais fatores terminariam por eliminar todas as espécies sobre sua influência de maneira totalmente aleatória, e isso obviamente não é o que acontece (ao menos na maioria das vezes), já que evidências apontam para uma seletividade taxonômica, funcional ou filogenética em relação a extinção ou ao seu risco associado [17, 28-31]. Inclusive o termo original “extinção seletiva” remonta a Wallace no seu artigo de 1855, “On the Law Which Has Regulated the Introduction of New Species” [32]. Essa seletividade só é possível porque espécies com características semelhantes respondem de maneira similar ao ambiente. Portanto, exceto em situações hipotéticas, fatores extrínsecos necessariamente interagem com características das espécies para gerar um risco real de extinção. Podemos chamar essas características ligadas às espécies de fatores intrínsecos [14, 17-19]. No caso do mamute o fato de ser grande e bem adaptado as condições de baixa temperatura o deram uma vantagem competitiva em relação as outras espécies durante a última glaciação. Porém, quando o clima esquentou não era vantajoso ser adaptado ao frio. Do mesmo modo, para os humanos o grande e peludo mamute era uma fonte preferencial de carne e pele. Em outras palavras, o mamute possuía uma série de fatores intrínsecos que eram desfavoráveis dado àquelas novas condições extrínsecas (esse processo de interação será discutido mais adiante).

Algumas características intrínsecas só emergem ao nível da espécie, como é o caso da distribuição geográfica e da abundância [33]. Diferente das outras características intrínsecas, como o tamanho do corpo ou a presença de pelos, a distribuição geográfica é impossível de ser obtida a partir de um único indivíduo (a não ser que este seja o único representante de uma espécie) e aparentemente não segue um modelo evolutivo [33]. Mais além de ser uma causa da extinção, essas características emergentes são propriedades da dinâmica surgimento-extinção, sendo comumente utilizadas para medir o risco associado ao

desaparecimento de uma espécie (quando = 0, há extinção) [34]. Portanto, podemos estabelecer uma categoria a parte para essas características.

Um outro fator importante é a dimensão evolutiva. Características como idade da espécie, taxa de diversificação do clado, distinção evolutiva, riqueza do clado ou mesmo o tipo de processo evolutivo que gerou aquela espécie, podem influenciar diretamente o seu próprio tamanho de distribuição geográfica ou abundância [35-38]. Essas características vão além da espécie, podendo apenas ser compreendidas em um contexto onde as relações evolutivas estão estabelecidas [33]. Também se diferenciam das outras características por sua posição temporal, já que os fatores evolutivos estão associados a extinção da espécie mesmo antes do seu surgimento, enquanto os fatores intrínsecos e extrínsecos estão associados ao período de tempo compreendido do surgimento à extinção das espécies.

Algumas causas de extinção só passam a ser relevantes quando a espécie tem sua distribuição restrita e uma baixa abundância. Entre esses fatores podemos incluir as flutuações estocásticas da população, deterioração genética e disfunção social (i.e quando não há indivíduos suficientes para a manutenção de um comportamento social da espécie) [39]. Quanto menor e mais isolada for uma espécie, mais susceptível ela estará a essas causas, as quais podemos chamar de fatores estocásticos [40]. Esses fatores se situam no tempo final da espécie, e foram considerados também como ultimate causes (causas últimas) da extinção [39].

Como ficou claro ao classificar esses grupos de causas/fatores que levam as espécies à extinção, uma outra propriedade emergiu ao longo do texto também como um efeito causal, o tempo. O tempo não é necessariamente independente das outras categorias (já que por exemplo ele depende dos fatores evolutivos ou da própria extinção, que é medida pelo tamanho populacional), mas é nele que a dinâmica entre as categorias é possível. Fatores intrínsecos e extrínsecos necessariamente variam ao longo do tempo, mudando também quais são as causas que estão afetando as espécies. Por exemplo, as mudanças climáticas só podem ser entendidas se observadas ao longo do tempo. O tempo também é responsável em parte pela dinâmica do tamanho da distribuição geográfica e da população, isso porque, assim como o espaço, o tempo determina os limites de expansão de uma espécie [41, 42]. Adicionalmente, por puro efeito da probabilidade, podemos afirmar que quanto mais tempo 17

uma espécie existe mais chance ela tem de sofrer com alguma causa que a leve a extinção [40, 43].

Adicionando complexidade: interações

Dificilmente uma única causa ou mesmo um grupo de causas (como as estabelecidas anteriormente) pode ser considerada como único fator direcionando a extinção. Ao contrário, o processo de extinção, assim como quase todos os processos ecológicos, normalmente envolve uma variedade de fatores e interações [18, 44-46], que por sua vez devem estar também variando ao longo do tempo e do espaço geográfico. Aceitar e entender essa complexidade é parte essencial para maximizar a eficiência das medidas proativas de conservação [47].

Em resumo, as interações aumentam o risco causado por fatores que isoladamente não causariam nenhuma ou pouca ameaça às espécies. Como é o caso do Batrachochytrium dendrobatidis. Essa espécie de fungo em determinadas condições climáticas dificilmente causa algum prejuízo a biodiversidade, mas quando se encontra em condições adequadas (normalmente relacionadas a grandes altitudes) ocasiona a quitridiomicose, doença que vem dizimando anfíbios no mundo inteiro [48]. Assim, podemos dizer que a presença do fungo e condições ambientais interagem para causar um maior risco às espécies. Nesse caso, fatores extrínsecos aparentemente inofensivos à biodiversidade interagem entre si para causar um alto risco de extinção. Do mesmo modo, fatores que isoladamente já proporcionariam algum risco, podem potencializar-se quando juntas, por exemplo, as mudanças climáticas poderão interagir com a fragmentação de florestas, a qual impedirá a dispersão de muitas espécies para áreas mais adequadas diante da variação do clima [49]. Pensar nessas interações nos permite tomar medidas de conservação como a construção de corredores ecológicos, que poderiam evitar essa última interação [50]. Ou como no caso da quitridiomicose, direcionar ações para evitar que o fungo chegue aos locais adequados à manifestação da patologia [48].

De maneira um pouco diferente, as interações entre fatores intrínsecos podem estar agindo em dois níveis. No primeiro nível, as interações entre variáveis devem gerar o que pode ser chamado de síndromes de caracteres [51] (ou response traits [52]). Para explicar melhor esse conceito, tomemos a capacidade de dispersão em aves como exemplo. Ela

depende do tamanho do corpo, taxas metabólicas, peso do esqueleto, aerodinâmica do corpo, tipo de penas, musculatura e muitas outras características [53]. É a interação entre essas variáveis que gera o que podemos medir como capacidade de dispersão. Por conseguinte, a dispersão não é algo que evolui diretamente na espécie, mas é resultado de uma série de outras variáveis, cada uma sofrendo suas próprias pressões evolutivas ao longo do tempo [52]. Em um segundo nível, as interações ocorrem da mesma maneira que em fatores extrínsecos. Por exemplo, uma espécie com maior capacidade de dispersão só poderá expandir-se latitudinalmente (diminuindo seu risco de extinção) se possuir uma série de outras características que irão determinar sua capacidade de tolerância ao frio ou ao calor.

Uma estratégia bastante usada nos estudos comparativos de extinção é uso de proxies, ou seja, caracteres das espécies que funcionam como substitutos para outras condições ecológicas, comportamentais, fisiológicas ou morfológicas [54, 55]. Essas variáveis normalmente são obtidas mais facilmente em comparação a outras, substituindo características que ainda não foram ou não podem ser observadas, como no caso de espécies fósseis. Uma variável comumente usada como proxy é o tamanho do corpo, já que é medida na maioria dos seres vivos e que em geral já está fornecida junto a descrição das espécies, além de servir como um substituto para diversas características ecológicas, como taxas metabólicas, estratégias reprodutivas, área de vida, etc. [56].

Porém, o uso de proxies deve também considerar as interações. Durante muito tempo foi estabelecido que espécies de tamanho de corpo maiores, por diversos motivos (mais detalhes no capítulo 1), são mais susceptíveis a extinção nos tempos atuais e em alguns momentos da história, como na extinção do cretáceo [30, 57]. Porém, as coisas não são tão simples. Estudos que incluíram a interação entre o tamanho do corpo e variáveis extrínsecas, mostraram que distintos fatores estão atuando ao longo do gradiente de tamanho corporal [18, 44, 45, 58]. Isso torna tudo mais complexo, principalmente porque como diferentes regiões apresentam diferentes condições, isso significa que ser maior ou menor pode ser bom ou ruim dependendo de onde a espécie estiver localizada, isto é, o efeito das interações entre fatores intrínsecos e extrínsecos podem levar a uma não estacionariedade espacial nas causas de extinção, tanto intraespecífico quanto interespecífico [19]. Além disso, para generalizar conclusões é preciso que o proxy funcione da mesma maneira entre diferentes grupos

taxonômicos/filogenéticos. Porém, dificilmente somos capazes de generalizar o efeito do tamanho do corpo (e de outros proxies) já que este reflete (substitui) diferentes aspectos em diferentes grupos. Por exemplo, o maior tamanho de corpo em mamíferos reflete maior susceptibilidade à caça [59], mas para uma serpente outras características que não o tamanho do corpo estão relacionadas a caça [60] (mais detalhes no capítulo 1). Finalmente, espécies de tamanho de corpo maior são menos susceptíveis aos fatores estocásticos quando comparados a espécies de tamanho menor, e como estes só passam a ser relevantes em baixas abundâncias, ou seja, próximo a sua extinção, podemos dizer que também o tempo influência o efeito do tamanho do corpo sobre o risco de extinção [61].

Essa natureza multivariada e de múltiplas interações relacionada as causas da extinção, que envolvem não só os fatores intrínsecos e extrínsecos como todos os outros, pode ser chamado de “caminhos da extinção” [62]. Mas como visto até agora, esses caminhos são multidimensionais, podendo variar na direção espacial, filogenética, ambiental, funcional e temporal. Se queremos entender os padrões gerais do risco de extinção, predições futuras ou extinções passadas deve-se estar atento a essa complexidade, afim de revelar os caminhos que levam uma espécie ao desaparecimento e evitar que interações prejudiciais às espécies ocorram [63].

Avaliando as causas na prática: o problema do viés de informação

Na prática, avaliar empiricamente as causas de extinção de uma espécie pode ser feito de duas maneiras, após o desaparecimento da espécie por meio dos registros fósseis, ou antes, utilizando dados de risco associado (normalmente sumarizadas em uma lista vermelha, sendo a mais ampla e conhecida a fornecida pela IUCN: http://www.iucnredlist.org/). Ambas as abordagens têm seus problemas e vantagens. Os fósseis representam eventos de extinções reais, porém recriar os fatores extrínsecos e intrínsecos dessas espécies é um desafio enorme e há uma dependência muito grande em torno dos proxies. Além disso, o processo de fossilização é em grande parte dependente das condições do ambiente e das características das espécies, o que pode causar uma serie de vieses ambientais, funcionais e filogenético nos estudos [64]. Por outro lado, utilizar espécies viventes se torna mais fácil uma vez que há mais informações e estas estão melhor distribuídas (em todas as dimensões) [65]. Todavia, o

valor de risco de extinção é uma medida fictícia baseado em informações limitadas, e por vezes esses problemas são desconsiderados.

A complexidade envolvida na questão da extinção limita a capacidade de utilizar experimentos, consequentemente, a grande maioria dos estudos é correlativa. Quando se trata de um estudo de uma ou poucas espécies, um desenho amostral bem feito e um processo adequado de obtenção de dados deve diminuir a maioria dos problemas, mas ao tentar elucidar padrões gerais utilizando muitas espécies é preciso estar atento a algumas dificuldades ligadas a informação. Essas dificuldades são causadas por dois problemas: (1) o processo de avaliação das espécies dificilmente segue um design aleatório e (2) nem sempre é possível atribuir um valor de risco às espécies avaliadas (como as espécies com “Dados Insuficiente”, DD). Então, há duas potenciais fontes de vieses em listas vermelhas direcionada pela disponibilidade de dados, a priori na própria seleção das espécies avaliadas ou gerado posteriormente pela distribuição das espécies DD.

Claramente, algumas regiões do mundo possuem mais informações que outras, ocasionadas ou pela facilidade de acesso, concentração de pesquisadores ou mesmo por questões históricas e políticas [66, 67]. Assim como algumas espécies são mais estudadas que outras, dado o número de especialistas, facilidade de estudo ou simplesmente carisma, o que finalmente pode gerar um viés taxonômico, filogenético (se os táxons estiverem mais relacionados evolutivamente) ou funcional (se os táxons compartilharem características funcionais) [67, 68]. Se as causas que determinam uma maior quantidade de dados sobre uma espécie também estão ligadas de alguma maneira ao seu risco de extinção, podemos estar fazendo generalizações sobre o processo de extinção que refletem apenas a nossa informação e não suas reais causas. De outro modo a extinção pode estar verdadeiramente ligada a informação, o que teria consequências negativas sobre nosso potencial conhecimento da biodiversidade, dado a atual taxa de perda de espécies.

Nossa informação sobre as espécies ainda é bastante dependente da classificação taxonômica, já que conectamos determinados dados ou conhecimento a identidade (ou ao nome) de uma espécie [69, 70]. O fato de que uma espécie só pode ter informações associadas ao ser formalmente descrita gera uma potencial correlação do conhecimento com o ano em que uma espécie foi descrita. Afinal, quanto mais tempo tivemos para estudar uma espécie 21

mais informações ela pode ter. Por fim, podemos imaginar que condições como tamanho da distribuição e tamanho da população, características utilizadas para classificar o risco de uma espécie, vão também estar ligados ao ano de descrição. Ou seja, todas essas conexões devem gerar um padrão temporal do risco de extinção, considerando o tempo como o ano em que uma espécie foi descrita. Ao mesmo tempo, isso pode acrescentar um fator de confusão, já que determinadas características que limitam a descoberta de uma espécie, também podem estar ligadas às espécies com pouca informação (como as espécies DD). Por exemplo, espécies pequenas, de coloração críptica, pouca atividade, hábitos noturnos, que vivem em locais de difícil acesso (etc.) devem ser muito mais difíceis de serem descritas e de serem estudadas, sendo por fim classificadas como DD. Estas espécies podem também estar ameaçadas, porém como estão numa categoria de incerteza em relação ao seu risco, suas características podem estar ficando de fora das grandes análises, levando-nos a concluir por exemplo que espécies pequenas são menos ameaçadas quando na verdade as espécies pequenas ameaçadas são consideradas DD.

Baseado na discussão feita até agora, proponho aqui uma imagem teórica do processo de extinção, suas causas envolvidas e como a nossa avaliação sobre o risco de extinção das espécies se relaciona a esses fatores (Figura 1). Esse esquema de relações ilustra a complexidade envolvida no entendimento da extinção e como nossa medida do risco de extinção é obtida indiretamente a partir de um filtro que é a quantidade de informação que temos disponíveis, a qual por sua vez é dependente de outros fatores também relacionados a extinção. Especificamente eu pretendo chamar a atenção para o processo temporal proposto no esquema. Nessa tese, juntamente com os respectivos coautores, discutimos dois desses aspectos temporais: (1) o tempo do surgimento até a extinção como um fator determinante no atual risco de extinção associado as espécies e o (2) tempo da descrição até os dias de hoje como um fator determinante da informação e consequentemente do nosso entendimento sobre o processo de extinção.

Figure 1. Esquema teórico representando as relações entres diferentes fatores que determinam o risco de extinção de uma espécie em uma representação na dimensão temporal. Nesse esquema é apresentado duas alternativas para obter o risco de extinção, como proposto teoricamente por meio da distribuição geográfica e abundância (seta ‘teórica’) e como é feito na prática (seta ‘prática’).

Apresentação da tese

A tese está dividida em quatro capítulos, cada um estruturado como um artigo científico seguindo as normas da revista ao qual foi ou será submetido, por isso o idioma utilizado é o inglês. Todavia, para essa tese, adaptamos algumas normas originais de cada revista a fim de melhorar a apresentação de cada capítulo para o leitor.

No primeiro capítulo, investigamos (junto aos coautores) como uma característica intrínseca, o tamanho do corpo, se relaciona ao processo de descrição e ao mesmo tempo ao risco de extinção em serpentes, criando uma conexão entre ambos. Mostramos como esses dois processos podem estar ligados por uma característica morfológica, já que ao mesmo tempo espécies menores são descritas mais recentemente e são classificadas como DD, criando um viés funcional nas análises de risco de extinção. Mais além, mostramos como por si só o ano de descrição está ligado diretamente ao risco de uma espécie, provavelmente por intermédio de ambas dependerem da distribuição geográfica e abundância das espécies.

No seguinte capítulo (2), mostramos que o padrão em que espécies mais ameaçadas são também descritas mais recentemente (encontrado em serpentes no capítulo 1), pode ser generalizado para todos os outros grupos de vertebrados e invertebrados que possuem seu risco avaliado. Mostramos que de fato a distribuição geográfica e abundância são os fatores responsáveis por essa relação. Assinalamos também as possíveis consequências negativas dessa relação sobre o nosso conhecimento acerca da biodiversidade e que medidas podem ser tomadas para minimizar esse impacto.

No capítulo 3 dessa tese, passamos a investigar o efeito dos fatores evolutivos sobre o risco de extinção das espécies de plantas. Avaliamos principalmente como o tempo em que uma espécie surgiu se correlaciona aos outros fatores evolutivos e por sua vez ao risco de extinção que é atualmente observado nas espécies. Mostramos que essa relação ocorre possivelmente mediada pela história biogeográfica do local, para tanto propomos um modelo teórico de como isso deve estar ocorrendo. Indicamos também, como a relação do risco com as características evolutivas pode gerar uma perda desproporcional de diversidade filogenética.

Por fim, no último capítulo, reunimos muitas dos aspectos metodológicos utilizados e desenvolvidos durante a tese (e fora desta) em um pacote de R, nomeado como uma homenagem ao Laboratório de Ecologia Teórica e Síntese onde foi desenvolvido, o letsR. Nele, disponibilizamos uma seria de ferramentas para facilitar a obtenção, manipulação e análise de dados macroecológicos. Exemplificamos sua funcionalidade ao avaliar o padrão espacial global de descrição dos anfíbios, mostrando como este depende de questões como o tamanho da distribuição, atividade humana e movimento histórico dos naturalistas.

Com isso esperamos que algumas das relações apresentadas na figura 1 sejam melhor esclarecidas e o papel do tempo na extinção melhor compreendido. Respondendo especificamente: (1) qual o papel do efeito do ano de descrição sobre nosso conhecimento das espécies e no entendimento do seu risco de extinção, e (2) qual o efeito do tempo de surgimento de uma espécie sobre seu risco de extinção atual.

Referencias

1. Faria FFdA (2010) Georges Cuvier e a instauração da paleontologia como ciência. Tese (doutorado) - Universidade Federal de Santa Catarina, Centro de Filosofia e Ciências Humanas, Programa de Pós-Graduação Interdisciplinar em Ciências Humanas, Florianópolis.

2. Cuvier G and Saint-Hilaire G (1795) Sur les espèces d’Éléphans. Bulletin de sciences par la Societé Philomathique de Paris 1: 90.

3. Rappaport R (1978) Geology and orthodoxy: the case of Noah's Flood in eighteenth- century thought. The British journal for the history of science 11: 1-18.

4. Nogués-Bravo D, Rodríguez J, Hortal J, Batra P and Araújo MB (2008) Climate change, humans, and the extinction of the woolly mammoth. PLoS Biol 6: e79.

5. Dirzo R, Young HS, Galetti M, Ceballos G, Isaac NJB, et al. (2014) Defaunation in the Anthropocene. Science 345: 401-406.

6. Vitousek PM, Mooney HA, Lubchenco J and Melillo JM (1997) Human Domination of Earth's Ecosystems. Science 277: 494-499.

7. Oreskes N (2004) The Scientific Consensus on Climate Change. Science 306: 1686- 1686.

8. Pimm SL, Jenkins CN, Abell R, Brooks TM, Gittleman JL, et al. (2014) The biodiversity of species and their rates of extinction, distribution, and protection. Science 344: 1246752.

9. Barnosky AD, Matzke N, Tomiya S, Wogan GO, Swartz B, et al. (2011) Has the Earth/'s sixth mass extinction already arrived? Nature 471: 51-57.

10. Ceballos G, Ehrlich PR, Barnosky AD, García A, Pringle RM, et al. (2015) Accelerated modern human–induced species losses: Entering the sixth mass extinction. Science Advances 1: e1400253. 25

11. Worm B, Barbier EB, Beaumont N, Duffy JE, Folke C, et al. (2006) Impacts of biodiversity loss on ocean ecosystem services. science 314: 787-790.

12. Mace GM, Norris K and Fitter AH (2012) Biodiversity and ecosystem services: a multilayered relationship. Trends Ecol Evol 27: 19-26.

13. Chapin III FS, Zavaleta ES, Eviner VT, Naylor RL, Vitousek PM, et al. (2000) Consequences of changing biodiversity. Nature 405: 234-242.

14. McKinney ML (1997) How do rare species avoid extinction? A paleontological view. In: W. E. Kunin and K. Gaston, editors. The biology of rarity: causes and consequences of rare-common differences. Springer Science & Business Media.

15. Darwin C and Wallace A (1858) On the tendency of species to form varieties; and on the perpetuation of varieties and species by natural means of selection. Journal of the proceedings of the Linnean Society of London Zoology 3: 45-62.

16. Raup DM (1994) The role of extinction in evolution. Proceedings of the National Academy of Sciences 91: 6758-6763.

17. Purvis A, Agapow P-M, Gittleman JL and Mace GM (2000) Nonrandom extinction and the loss of evolutionary history. Science 288: 328-330.

18. Cardillo M, Mace GM, Jones KE, Bielby J, Bininda-Emonds OR, et al. (2005) Multiple causes of high extinction risk in large mammal species. Science 309: 1239- 1241.

19. Fritz SA, Bininda‐Emonds OR and Purvis A (2009) Geographical variation in predictors of mammalian extinction risk: big is bad, but only in the tropics. Ecol Lett 12: 538-549.

20. Smith KG, Lips KR and Chase JM (2009) Selecting for extinction: nonrandom disease‐associated extinction homogenizes amphibian biotas. Ecol Lett 12: 1069- 1078.

21. Sinclair ARE, Pech RP, Dickman CR, Hik D, Mahon P, et al. (1998) Predicting Effects of Predation on Conservation of Endangered Prey. Conserv Biol 12: 564-575.

22. Vieira MC and Almeida‐Neto M (2015) A simple stochastic model for complex coextinctions in mutualistic networks: robustness decreases with connectance. Ecol Lett 18: 144-152.

23. Bellard C, Genovesi P and Jeschke J (2016) Global patterns in threats to vertebrates by biological invasions. Proc R Soc B. The Royal Society. pp. 20152454.

24. Tylianakis JM, Didham RK, Bascompte J and Wardle DA (2008) Global change and species interactions in terrestrial ecosystems. Ecol Lett 11: 1351-1363.

25. Barve N, Barve V, Jiménez-Valverde A, Lira-Noriega A, Maher SP, et al. (2011) The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecol Model 222: 1810-1819.

26. Karger DN, Cord AF, Kessler M, Kreft H, Kühn I, et al. (2016) Delineating probabilistic species pools in ecology and biogeography. Global Ecol Biogeogr.

27. Colwell RK and Lees DC (2000) The mid-domain effect: geometric constraints on the geography of species richness. Trends Ecol Evol 15: 70-76.

28. Fritz SA and Purvis A (2010) Selectivity in mammalian extinction risk and threat types: a new measure of phylogenetic signal strength in binary traits. Conserv Biol 24: 1042-1051.

29. Johnson CN, Delean S and Balmford A (2002) Phylogeny and the selectivity of extinction in Australian marsupials. Anim Conserv 5: 135-142.

30. McKinney ML (1997) Extinction vulnerability and selectivity: combining ecological and paleontological views. Annu Rev Ecol Syst: 495-516.

31. Friedman M (2009) Ecomorphological selectivity among marine teleost fishes during the end-Cretaceous extinction. Proceedings of the National Academy of Sciences 106: 5218-5223.

32. Fowler CW and MacMahon JA (1982) Selective extinction and speciation: their influence on the structure and functioning of communities and ecosystems. Am Nat: 480-498.

33. Jablonski D (2008) Species selection: theory and data. Annual Review of Ecology, Evolution, and Systematics: 501-524.

34. IUCN (2001) IUCN Red List categories and criteria: Version 3.1. IUCN Species Survival Commission, Gland, Switzerland.

35. Davies TJ, Smith GF, Bellstedt DU, Boatwright JS, Bytebier B, et al. (2011) Extinction risk and diversification are linked in a plant biodiversity hotspot. PLoS Biol 9: e1000620.

36. Yessoufou K, Daru BH and Davies TJ (2012) Phylogenetic patterns of extinction risk in the Eastern Arc ecosystems, an African biodiversity hotspot. PLoS ONE 7: e47082.

37. Birand A, Vose A and Gavrilets S (2012) Patterns of species ranges, speciation, and extinction. The American Naturalist 179: 1-21.

38. Wang S, Chen A, Fang J and Pacala SW (2013) Why abundant tropical tree species are phylogenetically old. Proceedings of the National Academy of Sciences 110: 16039-16043.

39. Simberloff D (1986) The proximate causes of extinction. Patterns and processes in the history of life. Springer. pp. 259-276.

40. Lande R (1993) Risks of population extinction from demographic and environmental stochasticity and random catastrophes. Am Nat: 911-927.

41. Gaston KJ (2003) The structure and dynamics of geographic ranges. Oxford University Press.

42. Gaston KJ and Blackburn TM (1997) Evolutionary age and risk of extinction in the global avifauna. Evol Ecol 11: 557-565.

43. Pearson PN (1998) Speciation and extinction asymmetries in paleontological phylogenies: evidence for evolutionary progress? Paleobiology 24: 305-335.

44. Owens IP and Bennett PM (2000) Ecological basis of extinction risk in birds: habitat loss versus human persecution and introduced predators. Proc Natl Acad Sci USA 97: 12144-12148.

45. Isaac NJ and Cowlishaw G (2004) How species respond to multiple extinction threats. Proc R Soc Lond, Ser B: Biol Sci 271: 1135-1141.

46. Lee TM and Jetz W (2011) Unravelling the structure of species extinction risk for predictive conservation science. Proceedings of the Royal Society B: Biological Sciences 278: 1329-1338.

47. Cardillo M and Meijaard E (2012) Are comparative studies of extinction risk useful for conservation? Trends Ecol Evol 27: 167-171.

48. Fisher MC, Garner TW and Walker SF (2009) Global emergence of Batrachochytrium dendrobatidis and amphibian chytridiomycosis in space, time, and host. Annu Rev Microbiol 63: 291-310.

49. Opdam P and Wascher D (2004) Climate change meets habitat fragmentation: linking landscape and biogeographical scale levels in research and conservation. Biol Conserv 117: 285-297.

50. Noss RF (2001) Beyond Kyoto: Forest Management in a Time of Rapid Climate Change

51. Después de Kyoto: Manejo Forestal en Tiempos de Cambio Climático Acelerado. Conserv Biol 15: 578-590.

52. Chapin FS, Autumn K and Pugnaire F (1993) Evolution of Suites of Traits in Response to Environmental Stress. The American Naturalist 142: S78-S92.

53. Díaz S, Purvis A, Cornelissen JH, Mace GM, Donoghue MJ, et al. (2013) Functional traits, the phylogeny of function, and ecosystem service vulnerability. Ecology and evolution 3: 2958-2975.

54. Piersma T, PÉRez-Tris J, Mouritsen H, Bauchinger ULF and Bairlein F (2005) Is There a “Migratory Syndrome” Common to All Migrant Birds? Ann N Y Acad Sci 1046: 282-293.

55. Morales-Castilla I, Matias MG, Gravel D and Araújo MB (2015) Inferring biotic interactions from proxies. Trends Ecol Evol 30: 347-356.

56. Sekar S (2012) A meta‐analysis of the traits affecting dispersal ability in butterflies: can wingspan be used as a proxy? J Anim Ecol 81: 174-184.

57. Peters RH (1986) The ecological implications of body size. Cambridge University Press.

58. Gaston KJ and Blackburn TM (1995) Birds, body size and the threat of extinction. Philos Trans R Soc Lond B Biol Sci 347: 205-212.

59. Price SA and Gittleman JL (2007) Hunting to extinction: biology and regional economy influence extinction risk and the impact of hunting in artiodactyls. Proceedings of the Royal Society of London B: Biological Sciences 274: 1845-1851.

60. Jerozolimski A and Peres CA (2003) Bringing home the biggest bacon: a cross-site analysis of the structure of hunter-kill profiles in Neotropical forests. Biol Conserv 111: 415-425.

61. da Nóbrega Alves RR and Pereira Filho GA (2006) Commercialization and use of snakes in North and Northeastern Brazil: implications for conservation and management. Vertebrate Conservation and Biodiversity. Springer. pp. 143-159.

62. Pimm SL, Jones HL and Diamond J (1988) On the Risk of Extinction. The American Naturalist 132: 757-785.

63. Davidson AD, Hamilton MJ, Boyer AG, Brown JH and Ceballos G (2009) Multiple ecological pathways to extinction in mammals. Proceedings of the National Academy of Sciences 106: 10702-10705.

64. Di Marco M, Cardillo M, Possingham HP, Wilson KA, Blomberg SP, et al. (2012) A novel approach for global mammal extinction risk reduction. Conservation Letters 5: 134-141.

65. Sansom RS, Gabbott SE and Purnell MA (2010) Non-random decay of chordate characters causes bias in fossil interpretation. Nature 463: 797-800.

66. Rodrigues AS, Pilgrim JD, Lamoreux JF, Hoffmann M and Brooks TM (2006) The value of the IUCN Red List for conservation. Trends Ecol Evol 21: 71-76.

67. Hortal J, Jiménez‐Valverde A, Gómez JF, Lobo JM and Baselga A (2008) Historical bias in biodiversity inventories affects the observed environmental niche of the species. Oikos 117: 847-858.

68. Ladle R and Hortal J (2013) Mapping species distributions: living with uncertainty. Frontiers of Biogeography 5.

69. Pyšek P, Richardson DM, Pergl J, Jarošík V, Sixtová Z, et al. (2008) Geographical and taxonomic biases in invasion ecology. Trends Ecol Evol 23: 237-244.

70. McNulty K (2016) Hominin Taxonomy and Phylogeny: What's In A Name. Nature Education Knowledge 7: 2.

71. Gotelli NJ (2004) A taxonomic wish–list for community ecology. Philosophical Transactions of the Royal Society B: Biological Sciences 359: 585-597.

Capítulo 1: Body size, extinction risk and knowledge bias in New World snakes

Bruno Vilela, Fabricio Villalobos, Miguel Ángel Rodríguez e Levi Carina Terribile

Artigo publicado na revista PLoS ONE em novembro de 2014 DOI: 10.1371/journal.pone.0113429

*Devido ao tamanho do Material Suplementar deste artigo (Supporting information), o mesmo deverá ser acessado online, seguindo o DOI do artigo ou o seguinte link: http://goo.gl/8G7Ekr

Abstract

Higher extinction vulnerability of large-bodied species is commonly reported for vertebrate groups. We focused on the New World’s Colubroidea snakes to investigate this pattern and its potential relationship with species’ description years, considering three broad IUCN Red

List risk categories (Threatened, Non-Threatened, and Data Deficient) as well as Not-

Evaluated species. We found a negative relationship between body size and description year, with large-bodied species being described earlier. Description year also varied among risk categories, with Non-Threatened species being described earlier than Threatened species and both species groups earlier than Data Deficient species. On average, Data Deficient species also presented smaller body sizes, while no size differences were detected between

Threatened and Non-Threatened species. So it seems that smaller body sizes are related with species detectability, thus potentially affecting both when a species is described (smaller species tend to be described more recently) as well as the amount of information gathered about it (Data Deficient species tend to be smaller). Our data also indicated that, if Data

Deficient species were to be categorized as Threatened in the future, snake body size and extinction risk would be negatively related, contrasting with the opposite pattern commonly observed in other vertebrate groups.

Key-words: Body size; Colubroidea; Conservation status; Description year; Red Lists;

Reptiles.

1. Introduction

A general attribute that can be easily measured in every species on earth is its body size.

Understanding how to make ecological predictions based on body size remains an important research goal in ecology since, in general, it is easier to measure morphological traits than to directly assess ecological aspects[1, 2]. For instance, a major question is whether ecological features correlated with body size can render any size class more or less susceptible to extinction. In other words, are small- or large-bodied species (or any size class in between) more vulnerable to extinction?

Analyses based on species’ conservation statuses, such as those included in the IUCN

Red List of Threatened Species [3], have shown that Threatened vertebrate species tend to be larger than Non-Threatened ones, at least for mammals [4, 5], birds [6, 7], frogs [8], and marine fishes [9]. Therefore, some ecological implications of body size may be causing large- bodied species to become more vulnerable to extinction. For instance, large-bodied species tend to have lower local abundances and, hence, are more likely to suffer from endogamy and loss of genetic diversity from stochastic processes [10]. Also, larger species are often more prone to be targeted by human hunters, fishermen or for general commercialization [9,

11, 12], or to suffer from habitat loss [4] owing to their greater energetic demands and their consequent need for larger home ranges [13]. Finally, large-bodied species typically reproduce at lower rates and have smaller litters than small-bodied species, thus being less able to recover from population decline [8, 14].

Generalizing the above explanations to all taxa is challenging owing to different or unknown body size-ecology relationships in poorly known taxa. For example, one could argue that large body size does not necessarily imply higher energetic demands because metabolic rate also depends on behavior [15], or that large-bodied species do not always 35

show low local abundances [e.g. 16]. Additionally, extinction risk related to large body size may be undermined by other species’ traits conveying less vulnerability to local threats, such as large geographic ranges and high dispersal capability [17]. Moreover, exceptions to the large species–higher vulnerability pattern have been found, for example, in Australian

Elapidae snakes [18], freshwater fishes [9], and fossil bivalves [19].

Another caveat in generalizing the large species-vulnerability pattern is its dependence on conservation status assessments based on available information. Red Lists provide summarized information on species’ threats, population trends, and conservation status, but are not free of biases as they depend on data availability [20, 21]. Thus, it is possible that small-bodied species are poorly studied owing to their more cryptic nature, narrow ranges and less charismatic appeal. Indeed, such characteristics of small-bodied species make them hard to discover and thus tend to be scientifically described more recently

[22], resulting in less time for their study and, possibly, for accumulating relevant information to establish their level of threat. Overall, this could result in more small-sized species being assigned a conservation status related to uncertainty (i.e. categorized as Data

Deficient) or not even being evaluated in comparison with larger species, which means that conclusions about the large species-vulnerability pattern could be misleading when based solely on species categorized as Threatened or Non-Threatened.

Here we focus on the New World’s snakes (superfamily Colubroidea) for which we first asked (1) how species distribute within major IUCN extinction risk categories across countries and (2) whether species’ description years – and consequently the time that has been available to accumulate biological information about them – are related with these categories and with body size. We also tested (3) if the large species-higher vulnerability pattern holds for these snakes, and (4) explored the extent by which the existence of species 36

categorized as Data Deficient (i.e. with inadequate information on their distributions and/or population statuses to assess their risk of extinction [3]) and Not-Evaluated (i.e. that are awaiting evaluation against the criteria [3]) could affect the existence of such pattern.

2. Material and Methods

2.1 The data

We generated a checklist of all extant New World snakes (superfamily Colubroidea) using the Reptile Database [23]. The final list comprised 1,240 species, their description years, and the countries where they occur; that is, 23 countries plus the Insular Caribbean countries that we treated as a single geographical unit (hereafter Caribbean Region) owing to their smaller sizes and poorer snake faunas. Conservation statuses of snakes were obtained from the IUCN

Red List of Threatened Species [24], with a total of 456 species in our database (36.8%) being included in extinction risk categories. We could not use the finer IUCN classification scheme owing to low species numbers in some categories. Therefore, we employed two broader risk categories classifying species as either Threatened (TE; 4.4%) or Non-

Threatened (NT; 25%), with the former including three finer IUCN risk levels (Critically

Endangered, Endangered and Vulnerable) and the latter two (Near-Threatened and Least

Concern). We also considered two additional categories comprising those species classified as Data Deficient (DD; 7.4%) by the IUCN, as well as species not included in the Red List but present in our database, which we termed as Not-Evaluated (NE; 63.2%).

We obtained Maximum Total Lengths in millimeters – a commonly reported body size measure for snakes [e.g. 25]– for 665 species in our database (53.6%) from Terribile et al. [26] for viperids and elapids, and from different sources for colubrids [27-30] (see details in Table S1). We log10-transformed all length measures prior to analysis [e.g. 31]. 37

2.2 Numerical methods

Our sample was not obtained following a random design but instead comprised the available data. This could have resulted in a taxonomically biased database unsuitable to generate robust conclusions regarding New World snakes. We thus first checked whether our data were equally distributed among families, subfamilies and genera by computing the taxonomic distinctiveness metric and comparing it with a random distribution [see 32, 33].

We found that our sample was more representative (i.e. contained more supraspecific taxa) than expected by chance (Δ+ = 79.445; p <0.001). Hence, we concluded that severe taxonomic biases were unlikely in our data.

Different species share different proportions of evolutionary history; thus whenever species are the sampling units in parametric statistical analysis (see below), the assumption of independence among observations cannot always be assumed (i.e. the closer the species in the phylogeny the more similar their characteristics might be [34]). Therefore, if phylogenetic autocorrelation were to be found in species’ description years, body sizes, or both, this should be taken into account in the analyses. To evaluate the presence of phylogenetic autocorrelation in these variables, we used the most comprehensive phylogenetic hypothesis for reptiles available until now [35] that comprises 288 (43.3%) species in our database. We completed this phylogeny with the remaining species whose genera were already represented in the phylogeny (271, 40.8%) including them as polytomies; whereas we excluded the remaining species (106, 15.9%) from this analysis.

Subsequently, we generated phylogenetic global Moran’s Is and Moran’s I correlograms (in this case using five phylogenetic distance classes with equal number of observations [36]) for both description years and body sizes to asses levels of phylogenetic 38

autocorrelation in these variables. We found virtually no phylogenetic autocorrelation in description years, neither globally (global Moran’s I = 0.05; p<0.001) nor in any distance class (Fig. S1), indicating that there is no need to control for phylogenetic relatedness in analyses involving this variable. In contrast, we found significant phylogenetic autocorrelation for body size (global Moran’s I = 0.19; p<0.001), particularly in the first distance class (Fig. S2), indicating that closely phylogenetically related species tend to have more similar body sizes. Accordingly, we used Phylogenetic eigenVector Regression (PVR) to generate a set of variables (phylogenetic eigenvectors) representing the phylogenetic relationships among snake species, and then regressed body size against these variables using an iterative search for the subset of eigenvectors that reduce the largest amount of autocorrelation in regression residuals [see 37]. Two eigenvectors (first and third) were selected for this procedure, which we used as covariables accounting for body size phylogenetic autocorrelation in our statistical analyses (see below), thus guaranteeing the statistical assumption of data independence [37].

We used one-way ANOVA to investigate differences in species’ description years among risk categories and, if found, we applied planned comparison tests to evaluate a potentially structured sequence of risk categories according to species’ description years.

Specifically, we asked (1) if Non-Threatened species — which are expected to be abundant and, hence, easier to find — were described earlier than Threatened ones (description year for NT > TE); and if species in these two groups were described earlier than either (2) Data

Deficient (TE+NT > DD) or (3) Not-Evaluated species (TE+NT > NE).

Regarding snake body size, we used multiple regression analysis to investigate relationships between species’ body size and description years, and then used one-way

ANCOVA to check for differences in body size among risk categories. We also applied 39

planned comparison tests to investigate (1) if Threatened species were larger than Non-

Threatened (TE > NT), and if species in these two groups were larger than either (2) Data

Deficient (TE+NT > DD) or (3) Not-Evaluated (TE+NT > NE) species.

Mean sizes of Threatened and Non-Threatened species may change in the future owing to potential designation of Data Deficient and Not-Evaluated species within these risk categories. In order to explore these possibilities, we generated two extreme scenarios: that all Data Deficient or Not-Evaluated species end up being classified as either (4) Threatened

(i.e. TE←[DD or NE] > NT) or (5) Non-Threatened (i.e. TE > NT←[DD or NE]). Although not realistic, these two extreme scenarios represent the largest effects that could be expected for the future changes on the large body-extinction risk pattern as a result of including poorly known groups within a risk category. In all the analyses involving body size, we used the two selected phylogenetic eigenvectors previously described as covariables to account for the phylogenetic autocorrelation existing in this variable (see above).

For all analyses, p-values were computed using simple F-tests as well as a randomization protocol that generated null distributions of body sizes and description years by randomly reshuffling these data across species 1,000 times. For each of these subsamples we obtained the F value of the corresponding analysis (i.e. ANOVA or regression), and the resulting 1,000 values were compared with the empirical F value. P-value results thus generated are expected to be more robust against potential biases in data [38], but in our case these were qualitatively similar to the ones given by classical F-tests, and we only reported the latter for simplicity. All analyses were performed in R 3.1.0[39], using the packages ape

[40], phytools [41] and letsR [42].

3. Results

3.1 Snake richness, extinction risks and geographic patterns

The IUCN evaluation of the conservation status of Colubroidea species across the New

World is far from complete (Fig. 1). This is less dramatic for North American countries

(Canada, USA and Mexico) and Chile (<30% of their species await evaluation) than for the other countries, which exhibited between 58% and 87% Not-Evaluated species. Mexico holds the richest Colubroidea snake fauna in the New World (368 species, 29.7%) but also the largest number of species in the Threatened (26.1%) and Data Deficient (71.7%) categories, despite its relatively higher evaluation completeness. These patterns convey on

Mexico a strategic importance for Colubroidea conservation. Brazil (334 species) and

Colombia (268 species) are also species-rich, and Brazil and the Caribbean Region rated second and third in Threatened species numbers (10 and 9, respectively), so these countries also deserve special attention.

3.2 Risk categories and species’ description years

We found significant differences in species’ description years among extinction risk categories (F3;661=20.65, p<0.001; Fig. 2), and in two of our planned comparison tests: NT >

TE (F=3.222; p=0.001) and TE+NT > DD (F=-5.169; p<0.001), but not in the third: TE +

NE > NE (F=-0.557; p=0.577). This suggests a relationship between extinction risk and the temporal sequence of species descriptions in which Non-Threatened species tend to be described earlier (23 years on average) than Threatened species, and species in these two groups earlier (55 years) than Data Deficient species, but not earlier than Not-Evaluated species.

3.3 Snake body size and extinction risk categories

Mean body length of New World snakes was 913 ± 21.6 mm, ranging from 150 mm for

Tantillita brevissima to 3600 mm for Lachesis muta. Regressing species body length against description year revealed a weak, but statistically significant negative relationship (r= -0.53; d.f. = 599; p<0.001; Fig. 3). That is, recently described species tend to be shorter than species described in previous years. Furthermore, significant differences were detected for body lengths across risk categories (F3;597=6.02, p<0.001; Fig. 4a), and for the planned comparison checking for smaller body size in DD species (i.e. TE+NT > DD, F= 3.575, p<0.001), but not for the other comparisons tested (TE+NT > NE: F= 0.015, p= 0.987, and TE > NT: F=-

0.334, p=0.738). In fact, the latter comparison would still remain non-significant even if all

DD species were to be classified as NT in the future (TE > NT←DD; F2; 597= 0.015; p =

0.901); whereas the same comparison would become significant if all DD species were to be classified as TE (TE←DD > NT; F2; 597 = 9.377, p = 0.002). That is, Threatened species would then be considered shorter, on average, than Non-Threatened species. Conversely, potential future evaluation of Not-Evaluated species seems unlikely to alter the mean sizes of Threatened and Non-Threatened species, as indicated by the lack of support found for the corresponding simulated scenarios (TE← NE > NT; F2;598 = 0.195, p = 0.659 and TE >

NT←NE; F2; 598 = 0.0535, p = 0.818).

We obtained similar results between the New World data and data based on the

Mexican species only. There were significant differences in mean body length across risk categories (F3; 231=10.63, p<0.001), and the only significant planned comparisons were

TE+NT > DD and TE←DD > NT (F=3.322, p<0.001; and F = 8.525, p = 0.003, respectively).

Therefore, differences across risk categories found using all New World data were not an artifact caused by the high concentration of Data Deficient species in Mexico (Fig. 1). 42

In sum, the data indicate that Data Deficient species were, on average, 358 and 381 mm shorter than Threatened and Non-Threatened species, respectively, and no differences in body length were detected either between Threatened and Non-Threatened snakes (although this could change in the future if current DD species were classified as TE), nor between these groups and Not-Evaluated snakes.

4. Discussion

Previous work on mammals [4, 5], birds [6, 7], frogs [8] and marine fishes [9] found a positive relationship between body size and extinction risk. However, our results for New

World Colubroidea snakes not only did not find this large species-higher vulnerability pattern, but indicated that if all species currently considered as Data Deficient were to be classified as Threatened in the future, the pattern would actually be the reverse (i.e. smaller species would be, on average, more at risk than larger species).

Body size has been commonly related with other biological traits affecting species’ vulnerability to threats [4, 8]. For example, in endotherms, body size is often correlated with metabolic rate [43], and higher metabolic rates often mean increased food intakes, hence larger home ranges, which in turn may lead to higher susceptibility to habitat loss, the main contemporary threat to biodiversity [44]. However, energetic demands in ectotherms are typically much lower than in endotherms [45]. Therefore, even if correlated with body size, metabolic rate in ectotherms may not vary drastically among species of different body size and even large-bodied species could potentially survive in small areas after habitat loss, if enough food and microhabitat remain available [43]. Thus, we may expect the impact of habitat loss to be detrimental for large-bodied snakes but not as strong for large-bodied endotherms. 43

Species’ geographic range size has also been related to extinction risk [46, 47] and it is usually assumed that large ranges confer less vulnerability to extinction as local threats are unlikely to strongly affect species’ total distributions [47]. In snakes, range size is also positively correlated with body size [25, 48]. Additionally, low reproductive rates have also been associated with increased vulnerability to extinction [8, 14]. However, in contrast to other vertebrates where reproductive rate correlates negatively with body size, large-bodied snakes often have higher reproductive rates than smaller snakes [49]. Higher reproductive rates may confer higher fitness to species, increasing population abundance and facilitating recovery from disturbance, making species less vulnerable to stochastic processes such as genetic drift and endogamy [10]. These aspects (i.e. the positive relationships of snake body size with range size and reproductive rate) may, at least partially, explain the lack of relationship between body size and extinction risk found in the present study.

Also, compared to mammal species, snakes present a narrower range of body size variation that could minimize differences between large and small species related to threat vulnerability. Additionally, unlike birds and mammals, commercial exploitation of snakes is usually linked to traits not strictly related to body size, such as magic rituals and skin commerce [50, 51]. Thus, we could expect that large-bodied snakes are not necessarily more prone to hunting like it is for mammals and fishes [4, 9]. Therefore, large and small snake species may be similarly affected by threats, hindering the observation of a positive relationship between body size and extinction risk.

Beyond body size, other traits may better predict snakes’ conservation status. For instance, Reed & Shine [18] found that behavioral characteristics are more important than body size in determining the conservation status of Australian Elapidae snakes. They suggested that sit-and-wait foragers are more likely to be threatened than active predators, 44

mainly because the former depend on specific habitats that are more susceptible to anthropogenic activities. Filippi & Luiselli [52] also found that some life-history traits (e.g. synchronized mating season) increased the risk of extinction of Italian snakes. These studies highlight the need to include other species’ traits, such as behavioral characteristics, to determine the relationship between different traits and extinction risk in snakes.

The year of description may affect the definition of species conservation status and related traits. For example, species described a long time ago have had more time to be studied and, consequently, more populations may have been discovered, enabling their categorization as less vulnerable or Non-Threatened. In our analyses, the year of description varied among risk categories. Non-Threatened species had the lowest mean description year, suggesting they had been easier to find, which is likely related to larger geographic ranges and/or abundances, two common characteristics of less vulnerable species. Thus, description year may determine the amount of available information, which in turn influences the categorization process. Accordingly, species classified as Non-Threatened have been described earlier, followed by Threatened species, and later by Data Deficient species.

Further research on other taxa may establish the generality of the temporal pattern of description year among risk categories across vertebrate groups. Finally, unlike Data

Deficient species, Not-Evaluated species did not show description year differences with either Threatened or Non-Threatened species; hence future risk categorization of species in this group are unlikely to modify our observed patterns.

To our knowledge, this is the first study relating vertebrate body size and extinction risk that has included Data Deficient species and, interestingly, we found that these species have, on average, smaller body size than species in categories with ‘adequate data’ (i.e.

Threatened and Non-Threatened snakes). This finding suggests that small-bodied snake 45

species have been poorly studied compared to large-bodied species. Such interpretation is consistent with our finding of a negative relationship between body size and description year, with a similar relationship found by Reed & Boback [53] for Australian and North American snakes, and with our observation that, on average, Data Deficient species have been the ones described more recently. New snake species are rarely encountered [54], possibly because snakes are typically hard to find in the field owing to their low activity rate, cryptic nature, and solitary habits. Our results clearly suggest that these challenges might be greater for smaller snakes, which in turn could be affecting the evaluation of the conservation status of species in this size class.

In summary, our data showed that the large species-higher vulnerability pattern present in different vertebrate taxa could not be generalized to New World snakes and suggests that care should be taken when trying to describe relationships between species traits and conservation status. For instance, comparing the body size between Threatened and Non-

Threatened species may hide the influence of poorly-studied species in determining observed patterns. Future studies should consider these possibilities, analyzing all species and taking care of potential biases in order to clearly understand what makes species more or less vulnerable to extinction and contribute to their conservation.

Acknowledgements

We thank B. Lima for discussions, S. Hereld for reviewing the English and the two anonymous reviewers for their helpful comments on the manuscript. B.V. was supported by a CAPES grant for doctoral studies, F.V. by a CNPq postdoctoral fellowship, L.C.T. by

CNPq (Process n. 473788/2009-8), and M.Á.R. by the Spanish Ministry of Economy and

Competitiveness (grant: CGL2010-22119). 46

References

1. Peters RH (1986) The ecological implications of body size. Cambridge University

Press.

2. Bonner JT (2011) Why size matters: from bacteria to blue whales. Princeton

University Press.

3. IUCN (2001) IUCN Red List categories and criteria: Version 3.1. IUCN Species

Survival Commission, Gland, Switzerland.

4. Cardillo M, Mace GM, Jones KE, Bielby J, Bininda-Emonds OR, et al. (2005)

Multiple causes of high extinction risk in large mammal species. Science 309: 1239-

1241.

5. Fritz SA, Bininda‐Emonds OR and Purvis A (2009) Geographical variation in

predictors of mammalian extinction risk: big is bad, but only in the tropics. Ecol Lett

12: 538-549.

6. Gaston KJ and Blackburn TM (1995) Birds, body size and the threat of extinction.

Philos Trans R Soc Lond B Biol Sci 347: 205-212.

7. Bennett PM and Owens IP (1997) Variation in extinction risk among birds: chance

or evolutionary predisposition? Proc R Soc Lond, Ser B: Biol Sci 264: 401-408.

8. Cooper N, Bielby J, Thomas GH and Purvis A (2008) Macroecology and extinction

risk correlates of frogs. Global Ecol Biogeogr 17: 211-221.

9. Olden JD, Hogan ZS and Zanden M (2007) Small fish, big fish, red fish, blue fish:

size‐biased extinction risk of the world's freshwater and marine fishes. Global Ecol

Biogeogr 16: 694-701.

10. Petchey OL and Belgrano A (2010) Body-size distributions and size-spectra:

universal indicators of ecological status? Biol Lett: rsbl20100240.

11. Owens IP and Bennett PM (2000) Ecological basis of extinction risk in birds: habitat

loss versus human persecution and introduced predators. Proc Natl Acad Sci USA 97:

12144-12148.

12. Jerozolimski A and Peres CA (2003) Bringing home the biggest bacon: a cross-site

analysis of the structure of hunter-kill profiles in Neotropical forests. Biol Conserv

111: 415-425.

13. McNab BK (1963) Bioenergetics and the determination of home range size. Am Nat:

133-140.

14. Cardillo M (2003) Biological determinants of extinction risk: why are smaller species

less vulnerable? Anim Conserv 6: 63-69.

15. Biro PA and Stamps JA (2010) Do consistent individual differences in metabolic rate

promote consistent individual differences in behavior? Trends Ecol Evol 25: 653-659.

16. Luiselli L, Akani GC, Rugiero L and Politano E (2005) Relationships between body

size, population abundance and niche characteristics in the communities of snakes

from three habitats in southern Nigeria. J Zool 265: 207-213.

17. Brown JH (1995) Macroecology, University of Chicago Press. Chicago, Illinois,

USA.

18. Reed RN and Shine R (2002) Lying in wait for extinction: ecological correlates of

conservation status among Australian elapid snakes. Conserv Biol 16: 451-461.

19. Harnik PG (2011) Direct and indirect effects of biological factors on extinction risk

in fossil bivalves. Proc Natl Acad Sci USA 108: 13594-13599.

20. Rodrigues AS, Pilgrim JD, Lamoreux JF, Hoffmann M and Brooks TM (2006) The

value of the IUCN Red List for conservation. Trends Ecol Evol 21: 71-76.

21. González-Suárez M, Lucas PM and Revilla E (2012) Biases in comparative analyses

of extinction risk: mind the gap. J Anim Ecol 81: 1211-1222.

22. Scheffers BR, Joppa LN, Pimm SL and Laurance WF (2012) What we know and

don’t know about Earth's missing biodiversity. Trends Ecol Evol 27: 501-510.

23. Uetz P and Hallermann J (2012) The reptile database. Zoological Museum, Hamburg,

Germany Electronic database accessible at http://wwwreptile-databaseorg Accessed

on 10 December 2012.

24. IUCN (2012) IUCN red list of threatened species. Version 2012.2. International

Union for the Conservation of Nature Gland, Switzerland. Electronic database

acessible at http://www.iucnredlist.org. Acessed on 09 December 2012.

25. Reed RN (2003) Interspecific patterns of species richness, geographic range size, and

body size among New World venomous snakes. Ecography 26: 107-117.

26. Terribile LC, Diniz‐Filho JAF, Rodríguez MÁ and Rangel TFL (2009) Richness

patterns, species distributions and the principle of extreme deconstruction. Global

Ecol Biogeogr 18: 123-136.

27. Boback SM and Guyer C (2003) Empirical evidence for an optimal body size in

snakes. Evolution 57: 345-451.

28. Freitas MA and Silva TFS (2005) A herpetofauna da Mata Atlântica nordestina.

USEB Pelotas, RS.

29. Freitas MA and Silva TFS (2007) A herpetofauna das caatingas e áreas de altitudes

do nordeste Brasileiro. USEB.

30. Köhler G (2003) Reptiles de Centroamérica. Herpeton. 49

31. Olalla‐Tárraga MÁ, Rodríguez MÁ and Hawkins BA (2006) Broad‐scale patterns of

body size in squamate reptiles of Europe and North America. J Biogeogr 33: 781-

793.

32. Clarke K and Warwick R (1999) The taxonomic distinctness measure of biodiversity:

weighting of step lengths between hierarchical levels. Mar Ecol Prog Ser 184: 21-29.

33. Clarke K and Warwick R (2001) A further biodiversity index applicable to species

lists: variation in taxonomic distinctness. Marine ecology Progress series 216: 265-

278.

34. Harvey PH and Pagel MD (1991) The comparative method in evolutionary biology.

Oxford university press Oxford.

35. Pyron RA, Burbrink FT and Wiens JJ (2013) A phylogeny and revised classification

of Squamata, including 4161 species of lizards and snakes. BMC Evol Biol 13: 93.

36. Diniz‐Filho JAF (2001) Phylogenetic autocorrelation under distinct evolutionary

processes. Evolution 55: 1104-1109.

37. Diniz‐Filho JAF, Bini LM, Rangel TF, Morales‐Castilla I, Olalla‐Tárraga MÁ, et al.

(2012) On the selection of phylogenetic eigenvectors for ecological analyses.

Ecography 35: 239-249.

38. Crowley PH (1992) Resampling methods for computation-intensive data analysis in

ecology and evolution. Annu Rev Ecol Syst: 405-447.

39. Team RC (2014) R: A Language and Environment for Statistical Computing. R

Foundation for Statistical Computing, Vienna, Austria, 2012. ISBN 3-900051-07-0.

40. Paradis E, Claude J and Strimmer K (2004) APE: analyses of phylogenetics and

evolution in R language. Bioinformatics 20: 289-290.

41. Revell LJ (2012) phytools: an R package for phylogenetic comparative biology (and

other things). Methods Ecol Evol 3: 217-223.

42. Vilela B and Villalobos F (2014) letsR: Tools for data handling and analysis in

macroecology. R package version 1.1.

43. Gillooly JF, Brown JH, West GB, Savage VM and Charnov EL (2001) Effects of size

and temperature on metabolic rate. Science 293: 2248-2251.

44. Dirzo R and Raven PH (2003) Global state of biodiversity and loss. Annual Review

of Environment and Resources 28: 137-167.

45. Pough FH (1983) Amphibians and reptiles as low-energy systems. Behavioral

energetics: the cost of survival in vertebrates: 141-188.

46. Harris G and Pimm SL (2008) Range size and extinction risk in forest birds. Conserv

Biol 22: 163-171.

47. Payne JL and Finnegan S (2007) The effect of geographic range on extinction risk

during background and mass extinction. Proc Natl Acad Sci USA 104: 10506-10511.

48. Terribile LC, Diniz‐Filho JAF, Lima‐Ribeiro MdS and Rodríguez M (2012)

Integrating phylogeny, environment and space to explore variation in

macroecological traits of Viperidae and Elapidae (Squamata: Serpentes). J Zool Syst

Evol Res 50: 202-209.

49. Shine R (2005) Life-history evolution in reptiles. Annu Rev Ecol, Evol Syst: 23-46.

50. Alves RRdN and Pereira-Filho GA (2007) Commercialization and use of snakes in

North and Northeastern Brazil: implications for conservation and management.

Vertebrate Conservation and Biodiversity. Springer. pp. 143-159.

51. Fitzgerald LA and Painter CW (2000) Rattlesnake commercialization: Long-term

trends, issues, and implications for conservation. Wildl Soc Bull: 235-253. 51

52. Filippi E and Luiselli L (2000) Status of the Italian snake fauna and assessment of

conservation threats. Biol Conserv 93: 219-225.

53. Reed RN and Boback SM (2002) Does body size predict dates of species description

among North American and Australian reptiles and amphibians? Global Ecol

Biogeogr 11: 41-47.

54. Pincheira-Donoso D, Bauer AM, Meiri S and Uetz P (2013) Global taxonomic

diversity of living reptiles. PLoS ONE 8: e59741.

Figure 1. Country proportions of New World Colubroidea snake species in four conservation status categories: TE: Threatened, NT: Non-Threatened, DD: Data Deficient and NE: Not- Evaluated. For each country, circle size is proportional to the number of species it contains (note that the circle for Chile is hard to observe owing to its very low number of species (5) compared to the other countries). 53

Figure 2. Description year comparison among four conservation status categories (TE: Threatened, NT: Non-Threatened, DD: Data Deficient and NE: Not-Evaluated) of New World Colubroidea snakes. Means and 95% Confidence Intervals are presented. Dashed line represents total mean. Note that the Y axis was drawn in log scale.

Figure 3. Relationship between description year and body size (measured as log10 maximum total body length) in New World Colubroidea snakes. Note that the Y axis was drawn in log scale.

Figure 4. Mean body size (log10 maximum total body length) values of Colubroidea snake species in four conservation status categories (TE: Threatened, NT: Non-Threatened, DD: Data Deficient and NE: Not-Evaluated) across the New World. Error bars are for 95% Confidence Intervals; dashed lines represent total means.

Capítulo 2: The loss of the unknown: or how higher extinction risks are biased towards newer species to science

Bruno Vilela, Fabricio Villalobos, Miguel Ángel Rodríguez e Levi Carina Terribile

Artigo preparado para a revista Conservation Letters

ABSTRACT

Human activities have accelerated species extinction rates, including the loss of unknown biodiversity and probably bringing up a new mass extinction. Here we investigated how such activities could impact our knowledge over biodiversity, by relating time of description, extinction risk and information availability. We gathered data on the conservation status and description year of 32,580 animals included in the IUCN global Red List. We show that the more recently described species are more likely to be more threatened. We provide evidence that this pattern is driven by the relationship between description year and specie’s geographic range and population size. Additionally, we show that recently described species are also the less-known, linking information and time. Together, these results indicate that current extinctions will have a higher impact on poorly known species and maybe even unknown ones, possibly driving a biodiversity information crisis.

Key-words: Animals; Data Deficient; Description year; Environmental Crisis; Invertebrates; IUCN; Red List; Threatened species; Time; Vertebrates.

INTRODUCTION

Human activities in the post-industrial era have produced, as a side effect, an environmental crisis comparable to the five largest extinction catastrophes in Earth’s history (Barnosky et al., 2011). This crisis affects all biodiversity on the planet, it has already caused the irreversible loss of many species and possibly uncountable others in the near future (Wearn et al., 2012; Ceballos et al., 2015; Urban, 2015). Scientists have been trying to address who are these species we are losing, where they are, and what we loose with them (e.g. Cardillo et al., 2005; Davidson et al., 2012; Pimm et al., 2014). However, the currently accelerated extinction rate may drive the loss of species we still know almost nothing about, maybe not even that they have ever existed.

Even though in the last century science advanced enormously, we are still far from a complete understanding of the world’s biodiversity, inevitably linked to the magnitude and complexity of this natural phenomenon. To illustrate our ignorance, only in the last five years, taxonomists described 16 new species of primates, meaning that more than 260 years since Carolus Linnaeus established the basis of the modern taxonomic classification were not enough to describe all the primates on Earth. If this is the case with our closest relatives, it is easy to imagine the size of our illiteracy related to other life forms that are much more diverse and far less studied (Mora et al., 2011; Scheffers et al., 2012; Costello et al., 2013). Nevertheless, this is not our only biodiversity knowledge gap. The so called Linnean shortfall, i.e. the incompleteness of our taxonomic knowledge, is only one between seven other shortfalls identified in a recent review by Hortal et al. (2015), which includes the Wallacean (distribution), Prestonian (abundance), Darwinian (evolutionary relationships), Eltonian (biotic interactions), Hutchinsonian (abiotic tolerances) and Raunkiæran (species traits). The extinction of partially or completely unknown species could imply the loss of unique pieces necessaries to solve these shortfalls. Hence, beyond the expected impacts on human economy and health, driven by the disruptive changes in ecosystem services (Chapin III et al., 2000), we may also be about to face a biodiversity knowledge crisis owing to the current extinction process.

The bad news is that a biodiversity knowledge crisis is exactly what is being expected to happen. Pimm et al. (2014) found that mammals, birds and amphibians evaluated by the

IUCN (International Union for Conservation of Nature) and described after 1900 showed a higher extinction rate and higher proportion of critically endangered species (i.e. at a very high risk of extinction in the wild) when compared to species described before such date. This indicates that recently described species are the ones disappearing at faster rates. In the same vein, studying the New World snakes, we showed that the mean description year (i.e. the year a species was formally described following the binomial nomenclature) in each of the conservation categories of IUCN red list increases towards the more threatened ones (Vilela et al., 2014). It was argued in both cases that this would occur mainly because species with large geographical range and/or greater population size should be easier to find. Consequently, such easily-found species would have more chance to be described, studied and protected earlier, and thus be considered less threatened than rare species (Pimm et al., 2014; Vilela et al., 2014).

Moreover, our knowledge about species and their description time seem to be related also. For instance, we found that the Data Deficient (DD) category (i.e. species without enough information to have its status established) presented, on average, the more recently described species (Vilela et al., 2014). This pattern was also previously noticed by Morais et al. (2013) in Brazilian tailless amphibians, when investigating the possible status of DD species. Therefore, if knowledge and description year are indeed related and if recently described species are more threatened, the species more vulnerable to extinction would be those for which science knows the less about. Indeed, in a recent assessment of herbarium orchid specimens of Madagascar, Roberts et al. (2016) found exactly that more recently described species are the ones with less information and smaller extent of occurrence, area of occupancy and abundance (as the number of specimens in the herbarium), which also puts at more risk of extinction.

Until now, most studies showing a relationship between species’ description year and conservation status or knowledge have done so for limited taxa/regions (Vilela et al., 2014; Roberts et al., 2016) or with restricted analysis (Pimm et al., 2014). Here, we tested whether this relationships holds (or not) for all animals evaluated by the IUCN Red List and having geographical range distribution available. More specifically, we addressed if (1) description year increases towards more threatened categories (description year vs. extinction) and (2) if DD species are the most recently described ones (description year vs. knowledge). 60

Additionally, we tested the hypothesis that (3) this pattern could be driven by the effect of recently described species having smaller geographic range and population sizes compared to species described earlier.

METHODS

Species data

We took advantage of the new letsR package (Vilela & Villalobos, 2015) to easily gather the description year and conservation status for 32,580 species of animals from the IUCN red list (IUCN, 2015). Note that from the ~80 thousands animals evaluated (by November 2015), we only included the species that had their correspondent spatial data available, which were manually downloaded from IUCN. The final list comprised the globally evaluated vertebrate classes of amphibians, mammals and birds, but also other taxa with less comprehensive data, including the reptiles, dragon flies, marine fishes, sea cucumbers, crayfish, shrimps, crabs and cone snails (genus Conus). The mammals were separated into terrestrial and marine species owing to their divergent geographic range sizes, which is markedly larger in marine mammals (Fig. S1).

For the statistical analysis, taxonomic groups were split into two groups, the More- Data (hereafter MD) group, including amphibians, birds, terrestrial mammals and reptiles, for which there was enough species representing each of the IUCN risk categories, and the Less-Data (hereafter LD) group, that includes the remaining taxonomic groups. For the former, the risk categories were transformed into Threatened [Vulnerable (VU), Endangered (EN) and Critically Endangered (CR)], Non-Threatened [Least Concern (LC) and Near Threatened (NT)] or kept as Data Deficient (DD). This transformation allowed the analysis of the LD group to have enough number of species in every category to reject the null hypothesis (H0: description year threatened = non-threatened). Species considered Extinct (EX) or Extinct in the Wild (EW) were a priori removed from the analysis.

Geographical range size vs. Description year

For each species, its geographic range size was calculated as the number of cells covered by the spatial polygons within a global grid of 1ºx1º resolution. This method was preferred as it required less computation time (essential given the high number of species involved) than directly calculating the polygons’ geometric area. This did not compromise the results, as both range size metrics are globally highly correlated (Smith et al., 1994; Orme et al., 2006). An ordinary linear regression was applied to check how much of the variation in the description year could be explained by the geographical range size.

Extinction risk vs. Description year

We used linear models to investigate the influence of species’ description year (predictor variable) on their assigned IUCN categories (response variable) in each taxa. To accomplish this, we transformed the IUCN categories of the MD group into an ordinal variable from 1 to 5 (LC = 1; NT = 2; VU = 3; EN = 4; CR =5), whereas in the LD group the extinction risk variable was kept as binary (i.e. threatened or not). For each taxa, we gathered the estimated beta coefficient generated from each of these models as the pattern’s effect size. To test if DD species were more recently described (description year vs. knowledge) than species in other categories, we applied a t-test, with the p-values corrected by the Holm-Bonferroni method for multiple comparisons (Holm, 1979).

Effect of geographical range and population size

IUCN specialists assign an extinction risk to species based on five different criteria related to their present or expected future conditions (IUCN, 2001). The criterion A is related to an observed or inferred population decline, B to the extent of occurrence or area of occupancy (two metrics of geographical range size, for a discussion see Gaston & Fuller, 2009), C and D to absolute population size of mature individuals and E to quantitative predictions of future extinction probability. If range size and/or population size is responsible for the description year vs. extinction risk pattern, then we would expect that as the proportion of species classified as threatened based on the criteria B and/or C/D increases in a group, the

description year vs. extinction risk pattern should be stronger. To test this, we applied an ANCOVA (Analysis of Covariance) to indicate how much variation of the pattern’s (description year vs. extinction risk) effect size is explained by the proportion of species classified as threatened based on the criteria B and/or C/D. The relationship between the criteria B/C/D proportion vs. pattern’s (description year vs. extinction risk) effect size shall also be mediate by how much geographical range size determines the description year (this also applies for the population size, however this data is not available for all the species). So, we included the coefficient of determination (R2) obtained from the regression between description year and range size in the ANCOVA analysis. Lastly, we also controlled for the amount of data information (i.e. LD and MD groups) by including this as covariable in the ANCOVA model.

RESULTS

In general, the relationship between description year and extinction risk was confirmed by our results. For the MD group, the average description year increased towards higher extinction categories in all taxa (Fig. 1; Table 1). Likewise, for the LD group, threatened species were in average more recently described than non-threatened species (Fig. 2; Table 1). The only exceptions were the marine mammals (F = 0.91; p = 0.36; DF = 62) and sea cucumbers (F = 0.54; p = 0.58; DF = 125), for which there was no significant differences in description years between threatened and non-threatened categories.

Our results also supported the hypothesis that description year and knowledge about species are associated. The DD category included the most recently described species in both MD and LD group (table 1: knowledge vs. description year). The only exceptions are the marine mammals (t = 1.59; p = 0.679) and crayfish (t = 2.17; p = 0.179), in which DD species did not present description years statistically different from the species in other categories.

Geographical range size was negatively associated with description year in all groups (Table 1), except the marine mammals (R2 = 0.03; p = 0.07). The amount of variation in description year explained by geographical range size ranged from 14.28% (in crabs) to 41.02% (in shrimps). In other words, recently described species tend to have a smaller

geographical range size, although this did not even explain half of the description year variation.

The ANCOVA model explained 93.83% (p < 0.001) of the description year vs. extinction risk pattern (Table 2). There is a general tendency for the groups in which the proportion used of criteria B (F = 3.77; p = 0.006) and/or C/D (F = 5.809; p < 0.001) to classify species as threatened were higher, to present a stronger effect size of the aforementioned pattern. Also, the description year vs. extinction risk pattern’s effect size increased (F = 4.948; p = 0.001) in taxonomic groups where geographical range size explained more of the description year variation.

DISCUSSION

Our findings showed that species’ description year is related to extinction risk and to our knowledge (information availability) about animal species. This reveals a general pattern in which recently described species are the ones more vulnerable to extinction and with less information. We demonstrated that this pattern is driven by the frequency in which criteria B, C and D (related to geographical range and population size) are used to classify species as threatened. This is also related to the fact that recently described species have narrower geographical range sizes. Indeed, Pimm et al. (2010) had already show that the description year of Brazilian flowering plants, amphibians, birds and mammals species was related to their range size, which together with the higher rates of deforestation in the region could indicate that recently described species are more threatened with extinction. All these evidences give support to the early ideas proposed by Pimm et al. (2014) and Vilela et al. (2014) that species’ geographical range and population size influence their description year and also extinction risk in animals, ultimately driving the observed pattern between these latter variables. This also seems to be the case for plants, as shown by Roberts et al. (2016), although more comprehensive studies are necessary to confirm it for this group.

Interestingly, Morais et al. (2013) advocated the possibility that recently described species could be misclassified as threatened. Because scientists still have a limited knowledge about geographical range and population size of these species, as more information comes up, estimations for both properties may increase (e.g. larger range or population size) and 64

species would later be classified as non-threatened. This potential misclassification could occur because the recent description of a species can be associated to other species characteristics such as small body size, cryptic behavior or its occurrence in regions of difficult access (Scheffers et al., 2012). If this affirmation holds, we would expect that early- described groups should present smaller effects sizes of the description year vs. extinction risk pattern, which was not true for our data (Fig. 1 and 2). We do not discard the outcome proposed by Morais et al. (2013), but it must be affecting fewer species and/or acting at smaller scales than expected. For example, geographical range and population size do not increase enough with the availability of more information to remove the species from a risk category.

One of the main concerns raised by our results is that, if the tendency of later described species being threatened of extinction can be extrapolated, DD species, that are the most recently described ones, are also more prone to be classified as threatened than what would be expected by the background extinction risk distribution. In other words, if a taxonomist describes a new species today, it has more chance to be threatened than would be for a species described 100 years ago. Moreover, species that are still unknown to science may follow the same fate. This supports a general statement that we may be about to lose species faster than we can name them (Costello et al., 2013). Consequently, this would lead the humanity into a biodiversity knowledge crisis.

Facing this pessimistic scenario, conservation policies must adapt to minimize such an imminent crisis. Knowing that species are not evolutionary independent from one another, we can expect that some specie’s characteristics must also represent, in some degree, redundant information. To protect threatened species carrying unique portions of evolutionary history (EDGE species: http://www.edgeofexistence.org/species/) seems to be a good strategy to maximize species evolutionary history representativeness, which typically implies also unique species’ traits and ecological aspects (Isaac et al., 2007). However, it will not be a good strategy when evolutionary distinctiveness (i.e. the amount of species’ unique evolutionary history when compared to others) is not correlated with other species characteristics. Thus, this EDGE approach must be expanded to include all aspects driving the uniqueness of species to assure that irreplaceable information will not be lost. In addition, conservationists can rely on different existing approaches to identify regions and direct future 65

surveys where unknown species are more likely to be discovered (Scheffers et al., 2012; Costello et al., 2013), for example using macroecological models (Diniz‐Filho et al., 2005), species distribution models (Pearson et al., 2007) or using proxies, as the number of DD species (Brito, 2010).

It is contradictory that current taxonomical practice is being undermine as a science when all life sciences still depend on the basic information associated to a species name (Gotelli, 2004; McNulty, 2016). Regardless of our capacity to confirm whether species are real entities or just exist for practical convenience (Rieseberg et al., 2006), it is clear that conservation biology strongly rely on it, as we can see in our results. Therefore, while radical changes do not happen in the way we approach conservation, it is still necessary to increase the description of species in other to increase information and our knowledge about them.

Losing species also represents losing millions of years of biological history. Like the burning of the Library of Alexandria, where innumerous ancient books where forever destroyed leading humanity to an unprecedented knowledge disaster, our results suggest that the sixth mass extinction will likely have a similar effect on our knowledge of biodiversity. Instead of books and scrolls, this time species are the ones being lost and perhaps even without no one reading their stories, silently disappearing into the unknown.

AKNOWLEDGEMENTS

We thank R. Dobrovolski and S. Gouveia for insightful comments on the ideas of the manuscript. We also thank E. Sideri for comments on grammar issues. BV thanks the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for a doctoral scholarship. FV was supported by a BJT “Science without Borders” grant from CNPq. LCT is supported by productivity grant from CNPq.

REFERENCES

Barnosky, A.D., Matzke, N., Tomiya, S., Wogan, G.O., Swartz, B., Quental, T.B., Marshall, C., McGuire, J.L., Lindsey, E.L. & Maguire, K.C. (2011) Has the Earth/'s sixth mass extinction already arrived? Nature, 471, 51-57.

Brito, D. (2010) Overcoming the Linnean shortfall: data deficiency and biological survey priorities. Basic and Applied Ecology, 11, 709-713.

Cardillo, M., Mace, G.M., Jones, K.E., Bielby, J., Bininda-Emonds, O.R., Sechrest, W., Orme, C.D.L. & Purvis, A. (2005) Multiple causes of high extinction risk in large mammal species. Science, 309, 1239-1241.

Ceballos, G., Ehrlich, P.R., Barnosky, A.D., García, A., Pringle, R.M. & Palmer, T.M. (2015) Accelerated modern human–induced species losses: Entering the sixth mass extinction. Science Advances, 1, e1400253.

Chapin III, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L., Hooper, D.U., Lavorel, S., Sala, O.E. & Hobbie, S.E. (2000) Consequences of changing biodiversity. Nature, 405, 234-242.

Costello, M.J., May, R.M. & Stork, N.E. (2013) Can we name Earth's species before they go extinct? science, 339, 413-416.

Davidson, A.D., Boyer, A.G., Kim, H., Pompa-Mansilla, S., Hamilton, M.J., Costa, D.P., Ceballos, G. & Brown, J.H. (2012) Drivers and hotspots of extinction risk in marine mammals. Proceedings of the National Academy of Sciences, 109, 3395-3400.

Diniz‐Filho, J.A.F., Bastos, R.P., Rangel, T.F., Bini, L.M., Carvalho, P. & Silva, R.J. (2005) Macroecological correlates and spatial patterns of anuran description dates in the Brazilian Cerrado. Global Ecology and Biogeography, 14, 469-477.

Gaston, K.J. & Fuller, R.A. (2009) The sizes of species’ geographic ranges. Journal of Applied Ecology, 46, 1-9.

Gotelli, N.J. (2004) A taxonomic wish–list for community ecology. Philosophical Transactions of the Royal Society B: Biological Sciences, 359, 585-597.

Holm, S. (1979) A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics, 65-70.

Hortal, J., de Bello, F., Diniz-Filho, J.A.F., Lewinsohn, T.M., Lobo, J.M. & Ladle, R.J. (2015) Seven shortfalls that beset large-scale knowledge on biodiversity. Annual Review of Ecology, Evolution, and Systematics, 46

Isaac, N.J., Turvey, S.T., Collen, B., Waterman, C. & Baillie, J.E. (2007) Mammals on the EDGE: conservation priorities based on threat and phylogeny. PLoS One, 2, e296.

IUCN (2001) IUCN Red List categories and criteria: Version 3.1. IUCN Species Survival Commission, Gland, Switzerland,

IUCN (2015) IUCN red list of threatened species. Version 2015.2. Acessed on 12 November 2015. In, http://www.iucnredlist.org. .

McNulty, K. (2016) Hominin Taxonomy and Phylogeny: What's In A Name. Nature Education Knowledge, 7, 2.

Mora, C., Tittensor, D.P., Adl, S., Simpson, A.G. & Worm, B. (2011) How many species are there on Earth and in the ocean?

Morais, A.R., Siqueira, M.N., Lemes, P., Maciel, N.M., De Marco, P. & Brito, D. (2013) Unraveling the conservation status of Data Deficient species. Biological conservation, 166, 98-102.

Orme, C.D.L., Davies, R.G., Olson, V.A., Thomas, G.H., Ding, T.-S., Rasmussen, P.C., Ridgely, R.S., Stattersfield, A.J., Bennett, P.M. & Owens, I.P. (2006) Global patterns of geographic range size in birds. PLoS Biol, 4, e208.

Pearson, R.G., Raxworthy, C.J., Nakamura, M. & Townsend Peterson, A. (2007) Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. Journal of Biogeography, 34, 102-117.

Pimm, S.L., Jenkins, C.N., Joppa, L.N., Roberts, D.L. & Russell, G.J. (2010) How many endangered species remain to be discovered in Brazil. Natureza & Conservação, 8, 71-77.

Pimm, S.L., Jenkins, C.N., Abell, R., Brooks, T.M., Gittleman, J.L., Joppa, L.N., Raven, P.H., Roberts, C.M. & Sexton, J.O. (2014) The biodiversity of species and their rates of extinction, distribution, and protection. Science, 344, 1246752.

Rieseberg, L.H., Wood, T.E. & Baack, E.J. (2006) The nature of plant species. Nature, 440, 524-527.

Roberts, D.L., Taylor, L. & Joppa, L.N. (2016) Threatened or Data Deficient: assessing the conservation status of poorly known species. Diversity and Distributions, n/a-n/a.

Scheffers, B.R., Joppa, L.N., Pimm, S.L. & Laurance, W.F. (2012) What we know and don’t know about Earth's missing biodiversity. Trends in ecology & evolution, 27, 501-510.

Smith, F.D., May, R.M. & Harvey, P.H. (1994) Geographical ranges of Australian mammals. Journal of Animal Ecology, 441-450.

Urban, M.C. (2015) Accelerating extinction risk from climate change. Science, 348, 571- 573.

Vilela, B. & Villalobos, F. (2015) letsR: a new R package for data handling and analysis in macroecology. Methods in Ecology and Evolution, 6, 1229-1234.

Vilela, B., Villalobos, F., Rodríguez, M.Á. & Terribile, L.C. (2014) Body Size, Extinction Risk and Knowledge Bias in New World Snakes. Plos ONE, 9, e113429.

Wearn, O.R., Reuman, D.C. & Ewers, R.M. (2012) Extinction debt and windows of conservation opportunity in the Brazilian Amazon. Science, 337, 228-232.

Table 1. Summary results of the: linear model testing extinction risk vs. description year per

taxa; linear regression between range size and description year; proportion of species

classified in any threaten category by the criteria B and C/D; and, t-test comparing DD

species’ description year against other categories (knowledge vs. description year), p-values

were corrected by a bonferroni-holm method.

Extinction risk vs. Range size vs. Proportion (%) Knowledge vs. Description year Description year of criteria: Description year p- t p-value Group Taxon Effect Size F DF R2 p-value B C/D value MD Amphibians 13.50 26.79 4739 <0.001 0.365 <0.001 78.90 10.96 -24.47 <0.001 Reptiles 14.91 17.56 3183 <0.001 0.374 <0.001 72.21 14.33 -15.00 <0.001 Mammals 11.58 16.85 4536 <0.001 0.362 <0.001 49.78 20.83 -20.18 <0.001 (terrestrial) Birds 14.49 29.36 9863 <0.001 0.303 <0.001 21.66 50.16 -12.12 <0.001 LD Mammals 14.42 0.91 62 0.364 0.031 0.071 0 52.94 -1.60 0.679 (marine) Odonata 38.44 6.95 1108 <0.001 0.342 <0.001 69.77 25.58 -13.42 <0.001 Marinefish 25.31 3.24 1523 0.001 0.274 <0.001 25.00 44.05 -6.82 <0.001 Sea 7.27 0.54 125 0.588 0.241 <0.001 0 18.75 -5.97 <0.001 cucumbers Crayfish 23.54 4.12 379 <0.001 0.306 <0.001 83.72 11.63 -2.18 0.179 Shrimps 47.76 9.32 466 <0.001 0.410 <0.001 33.94 63.30 -8.38 <0.001 Crabs 26.97 6.59 650 <0.001 0.143 <0.001 66.18 51.96 -5.50 <0.001 Cone snails 33.87 2.52 543 0.012 0.366 <0.001 58.70 41.30 -7.18 <0.001

Table 2. ANCOVA model (R2=0.9383; p <0.001) applied to explain the effect size of the relationship Extinction risk vs. Description year.

Predictor variables Effect size F p-value Proportion of criteria B 18.35 3.77 0.007 Proportion of criteria C/D 43.73 5.81 0.001 Description year vs. Range size (R2) 63.22 4.95 0.002 Groups 15.16 5.67 0.001

Figure 1. Description year comparison among the conservation status categories (DD: Data

Deficient; CR: Critically Endangered; EN: Endangered; VU: Vulnerable; NT: Near

Threatened; LC: Least Concern) of amphibians, reptiles, birds and terrestrial mammals (MD group). Means and 95% Confidence Intervals are presented.

Figure 2. Description year comparison among Threatened (including the Critically

Endangered, Endangered and Vulnerable species), Non-Threatened (including Near

Threatened and Least Concern species) and Data Deficient (DD) of marine mammals, dragon flies, marine fishes, sea cucumbers, crayfish, shrimps, crabs and conus (LD group). Means and 95% Confidence Intervals are presented.

Figure S1. Mammals geographic range size range frequency distribution (measured as the proportion of individuals per category), divided into terrestrial (solid gray bars) and marine

(striped black bars).

Capítulo 3: Biogeography underlies the effects of evolutionary history on current extinction risks in the Iberian flora

Bruno Vilela, Rafael Molina‐Venegas, Juan Carlos Moreno Saiz, Levi Carina Terribile e Miguel Á. Rodríguez

Artigo preparado para a revista Diversity and Distribution

*Devido ao tamanho do Material Suplementar deste artigo (Supporting information), o mesmo deverá ser acessado online, seguindo o seguinte link: http://1drv.ms/1mwWLz3

(A) ABSTRACT (B) Aim Exacerbated current extinction rates are largely caused by recent human activities. However, species are old entities for which understanding processes leading to their extinction requires taking into account their evolutionary history. Here we investigated whether and how vulnerability to extinction and plant species evolutionary history relate in Eastern Andalusia and the whole Spanish Iberian Peninsula. (B) Location Spanish Iberian Peninsula and Eastern Andalusia (B) Methods We combined existing IUCN-based red list regional assessments independently conducted for these regions with a phylogeny including 94% of Iberian plant genera to estimate: (1) phylogenetic signal in within-genus extinction vulnerabilities as proportions of per-genus of threatened species; (2) the relationships between extinction risk and four phylogenetic characteristics (taxon age, diversification rate, evolutionary distinctiveness and clade taxonomic richness, and within-genus extinction vulnerabilities) using a multiple GLM; (3) the expected phylogenetic diversity loss under a extinction model considering the current scenario of species vulnerability. (B) Results We found that extinction risk was phylogenetically aggregated and that older and evolutionarily distinct genera tend to be more vulnerable to extinction in Eastern Andalusia, but not in the Spanish Iberian Peninsula. The extinction model points to a disproportional loss of evolutionary history in Eastern Andalusia flora. Our data also uphold previous findings suggesting the existence of “ecological phantoms” in the Iberian flora; i.e., of older species carrying traits that reflect adaptations to ancient but no longer existing environments. This suggests that the biogeographic history acts as an agent of the observed patterns. (B) Main conclusions As a whole, our analysis supports a connection between plant extinction risk and evolutionary history mediated by biogeography, particularly for older and more evolutionarily distinct clades, with potentially negative consequences to the phylogenetic conservation of Eastern Andalusia region.

Key-words: Andalusia; Conservation Status; Diversification Rate; Evolutionary Distinctiveness; Phylogenetic Signal; Phylogenetic Diversity; Red Lists; Taxon Age; Vascular Plants.

(A) INTRODUCTION

The extinction process has a key role on species evolution, almost certainly appearing with the first living beings (Raup, 1994). This process has been so ostensible along life’s history, that current estimations uphold that at least 99% of all species that have ever existed ended up extinct (McKinney, 1997). Such extinctions are normally balanced by speciation, avoiding that all biodiversity vanishes (Quental & Marshall, 2013). Yet, five times in earth history, this balance was altered by stochastic events, increasing extinction rates and consequently, critically shrinking earth’s biodiversity (i.e. the five mass extinctions events; Raup & Sepkoski, 1982). Currently, evaluations indicate that a new mass extinction must be going on caused by human activities (McKinney & Lockwood, 1999; Barnosky et al., 2011; Alroy, 2015). Although this certainly is the main cause of Anthropocene extinctions, considering only our current temporal lapse to investigate such causes disregards, a priori, that species are historically dynamic, that extinction can be a continuous process in time and that extinction is, also, a natural process.

Long before humans appeared as a major threat to biodiversity, natural causes of species extinctions included, among others, natural disasters (e.g. Brusatte et al., 2015), change in environmental conditions (e.g. Finnegan et al., 2012), habitat loss (Fahrig, 1997), predation (Sinclair et al., 1998), diseases (e.g. Smith et al., 2009), competition exclusion (e.g. Jaeger, 1970) and mutualistic coextinctions (Vieira & Almeida‐Neto, 2015). These causes would negatively affect species’ extent of occurrence (i.e. the geographical range size of a species), area of occupancy (i.e. the total area actually occupied by a species inside its geographical range) and effective population size (i.e. the number of reproductive individuals representing the same allele distribution of the total population), all characteristics currently used to establish species’ conservation statuses (details on IUCN, 2001). Once reduced to small and geographically isolated populations, species would be vulnerable to ultimate causes of extinction such as demographic stochasticity, genetic deterioration, social dysfunction and extrinsic forces (Simberloff, 1986).

Alongside with natural causes, it is expected that evolutionary history can also affect species long-term persistence. This because there are enough evidence that links evolutionary history with direct measures of species extinction risk, like the species’ geographical range

size, area of occupancy and population size (e.g. Birand et al., 2012; Wang et al., 2013). The last decade brought considerable support to such statement benefiting from the increasing availability of high-resolution phylogenies and methodological advance (Purvis, 2008), which allowed the test of explicit hypothesis connecting phylogenetic traits (hereafter defined as any variable that is derived from a phylogeny) and current extinction risk (e.g. Gaston & Blackburn, 1997; Johnson et al., 2002; Meijaard et al., 2008; Safi & Pettorelli, 2010; Davies et al., 2011; Yessoufou et al., 2012; Arregoitia et al., 2013), therefore, considering the evolutionary history in analyses extinction risk is essential.

The most common and intuitive phylogenetic trait related to extinction risk is the age of a clade. Most of the evidences indicate a pattern relating higher vulnerability towards old clades in some taxa of mammals, birds and plants (McDowall, 1969; Hodgson, 1986; Gaston & Blackburn, 1997; Johnson et al., 2002; Meijaard et al., 2008; Daru et al., 2013). Although the mechanisms leading to this observed pattern were not directly tested, many hypotheses were proposed to explain it (see Table S1). In opposition, further investigations indicate that the pattern of old species being more vulnerable to extinction is not a general rule and can only be applied to specific clades and regions, because when investigated at coarse scales, the pattern disappears (Schwartz, 1993; Vamosi & Wilson, 2008; Arregoitia et al., 2013). Moreover, for the world primates (Arregoitia et al., 2013) and the cape flora (Davies et al., 2011) the inverse pattern was found, as recently evolved (younger) clades were at more risk than early evolved ones.

Still controversial, the age related extinction risk pattern customarily exhibits a correlation with other phylogenetic traits patterns. Studies with animals (mammals and birds) that associated early evolved clades to high extinction risk usually found such clades to be, also, species poor or highly evolutionary distinct (i.e. that contains in itself a higher amount of unique evolutionary history in the given context) (McDowall, 1969; Gaston & Blackburn, 1997; Johnson et al., 2002; Meijaard et al., 2008). This may indicate that the process that makes a clade more vulnerable to extinction could be also determining the disappearances of other species belonging to this same clade (generating the species poor condition), and in some cases, even pruned entire sister lineages (making the target clade more evolutionary distinct). Yet, we cannot discard the alternative that perhaps these are consequences of long- term lower speciation rates. On the other hand, in the cape flora (South Africa), Davies et al. 78

(2011) showed that pronounced diversification rate (i.e. the balance of speciation/extinction in a specific clade given its time of existence) was associated to higher species vulnerability, what finally led to higher rates of both speciation and extinction in younger and species rich clades. Davies et al. (2011) argued that such different patterns between animals and plants could be due to their different dominant speciation mode, as in plants the predominance of peripatric speciation (i.e. speciation in new and isolated geographical ranges by a small population) would more frequently generate new species with smaller geographic range sizes than under other models of speciation. However, the explanation does not seem to hold, because for primates the same pattern of younger clades more threatened was found by Arregoitia et al. (2013), while Daru et al. (2013) showed that evolutionary old and distinct clades of world mangroves were more threatened.

Independent of the causes driving the different patterns, there seems to be two cohesive scenarios when evolutionary history underlies the extinction process, where higher threatened clades tend to be: (1) evolutionary older, species poor, with lower diversification rate and representing large amount of unique evolutionary history (hereafter type 1 pattern); or (2) evolutionary younger, species rich, with higher diversification rate and representing limited amount of unique evolutionary history (hereafter type 2 pattern). Both scenarios can drive to a phylogenetic signal (i.e. the tendency of phylogenetically closer species to have similar characteristics) of extinction risk (Fritz & Purvis, 2010), even in the absence of other traits (not phylogenetic traits) driving it (which seems to be the case of plants Davies et al., 2011). In conservation terms, this could point toward a potential loss of specific regions from the tree of life, although the presence of a phylogenetic signal in extinction risk alone does not implicates a disproportional loss of evolutionary history (for a complete revision see Veron et al., 2015). The impact of the current extinction process on the evolutionary history can be expected to be different depending on the observed patterns described above. In the type 1 pattern (were older clades are more threatened), a combination of phylogenetically aggregated extinction risk and more risk towards evolutionary distinct clades can cause a greater loss of clades representing unique evolutionary history, consequently occasioning a larger proportion of evolutionary history loss (Davies, 2015). While in the type 2 pattern, the inverse combination would induce a minor proportion of evolutionary history loss (Davies, 2015).

In plants, the global evaluation carried out by Vamosi & Wilson (2008) identified that species-poor families are the most threatened. Although no pattern relating clade age or evolutionary distinctiveness with extinction risk were found, a disproportional amount of evolutionary history (when compared to random expectations) was predicted to be lost under future extinctions models – this disproportional loss is also globally expected in mammals (Huang et al., 2011) and birds (Purvis et al., 2000). However, contrary to birds and mammals, global analysis for plants are still very limited, as only 17% (20187 spp. on 12 November 2015), from the more of 350 thousand recognized plants on earth (TPL, 2013) were globally evaluated until now (IUCN, 2015). In this case, regional evaluations with good phylogenetic and conservation statuses data, allow more detailed and feasible conclusions. Such information is urgently needed in the face of the current extinction scenarios, mainly for plants, were evolutionary history seems to have an important role (Davies et al., 2011; Yessoufou et al., 2012; Daru et al., 2013).

Here we took advantage of two independent regional assessments of extinction risk for plants, one considering the Spanish Iberian Peninsula (Saiz et al., 2015) and the other in a Mediterranean sub region of the former (Blanca et al., 2009), to investigate the role of evolutionary history on the current extinction risk distribution in both scales. Specifically, we aim to: (I) indicate if the extinction risk is phylogenetic clustered; (II) identify whether clade age, evolutionary distinctiveness, clade species richness and diversification rate can affect the flora persistence; and (III) assess if the patterns found in the first and second objectives would lead to a disproportional loss of evolutionary history. We expected that if the proposition of Davies et al. (2011) – i.e. peripatric speciation mode more common in plants, causing new species small-range pattern – is applied in the region, a type 2 pattern of extinction risk will be found in both scales. Alternatively, we can expect that differences in the biogeographic history of the regions, species composition and scales will conduct to different type of extinction pattern. The absence of a pattern would indicate other causes beyond evolutionary history or traits connected to it driving the extinction risk of the regions. Concerning the third objective, we expect a disproportional loss under a type 1 pattern and the contrary under a type 2 pattern. In the absence of patterns we expect the extinctions to be equal to a random expectation.

(A) METHODS

(B) Species list and extinction risk data

We used two independent assessments of extinction risk, both based on the regional criteria of the International Union for Conservation of Nature (IUCN, 2010). The first list includes 6222 species of vascular plants from the Iberian Peninsula, excluding Portugal (Saiz et al., 2015). The second list includes 3255 species from Eastern Andalusia region (Blanca et al., 2009) (Fig. 1; for a short description of both areas see Appendix S2). We upscaled the lists for the genus level by calculating the proportion of threatened species in each genus. For this purpose, we considered all species as either Threatened (Critically Endangered, Endangered and Vulnerable) or Non-Threatened (Near-Threatened and Least Concern). We also added to the Threatened category the species classified as Extinct, Extinct in the Wild and Regionally Extinct. To avoid a bigger problem of information bias raised by Daru et al. (2013; see Table S1), we only considered a species as threatened if there was enough evidence to do so. Hence, in this work, species in the category Data Deficient (i.e. species without enough information to have its status confirmed) and Not Evaluated (i.e. species that were not included in the conservation assessment) were considered as Non-Threatened (see figures 2 and 5 of the Appendix S1). We solved all nomenclature conflicts between the two assessments following the list provided by Saiz et al. (2015), as this is the most recent and comprehensive for the region. The final lists include 1052 genus in the Iberian Peninsula and 843 genus in Eastern Andalusia with their correspondent proportion of threatened species.

(B) Phylogeny

We adopted the molecular phylogenetic hypothesis provided by Hinchliff & Smith (2014), constructed from a supermatrix using the available sequences from GenBank. The complete phylogeny comprises 13,093 genera of world plants, from which we made a subset tree to include only the genera contained in both lists. We changed all nomenclature conflicts between the phylogeny used and our conservation statuses data to fit the nomenclature provided in the original species list from Saiz et al. (2015). To confirm whether the analysis and, consequently, its results could be affected by the selected phylogenetic hypothesis, we repeated all the analysis using a specific phylogeny developed for the region, which includes 81

fewer genera (Molina‐Venegas & Roquet, 2014). The results remained qualitatively the same, so, for simplicity, we keep only the results and analysis using the proposed phylogeny of Hinchliff & Smith (2014) that embraces more genera.

(B) Phylogenetic signal

We explored different metrics to evaluate if there was a phylogenetic signal on the proportion of threatened species in each genus, following the suggestions of Münkemüller et al. (2012) and Diniz-Filho et al. (2012b). We first applied two metrics of autocorrelation that do not consider an evolutionary model (the so called “statistic” metrics), the global Moran’s I

(Moran, 1950) and Abouheif’s Cmean (Abouheif, 1999). These two metrics differ mainly on the pairwise distance matrix used, as Morans’ I can be applied with any distance matrix, while Abouheif’s Cmean uses a matrix based on topological distances (i.e. do not consider the branch lengths) with a diagonal zero (Pavoine et al., 2008). The following formula is provided by Gittleman & Kot (1990) to calculate the Global Moran’s I:

� � � ∑�=� ∑�=� �� (�� − �̅)(�� − �̅) � = � � � � ∑�=� ∑�=� �� ∑�=�(�� − �̅)

Where wij represents the phylogenetic distance between the species i and j, x represents the variable for which the autocorrelation is measured and n the total number of samples. The resulting value (for both metrics) generally range between 0-1 (in some specific cases it can goes beyond 1), where 0 is the absence of autocorrelation and 1 means that the phylogenetic distances can explain all the variable variation. The signal (+ or -) of the result indicates if the autocorrelation is positive (i.e. phylogenetic closer species tend to have more similar values) or negative (i.e. phylogenetic closer species tend to have more different values). We also explored the phylogenetic autocorrelation using a Moran’s I correlogram, which calculates the Global Moran’s I for different classes of phylogenetic distance, allowing a better understanding of the autocorrelation pattern. We set 16 equidistant phylogenetic classes and calculate the Moran’s I for each one of them.

Additionally, we applied two other metrics of phylogenetic signal that assume an evolutionary model (“model-based” metrics), Pagel’s λ (Pagel, 1999) and Blombergs’s K

(Blomberg et al., 2003). The Pagel’s λ is originally a metric of tree transformation in which a value of λ=0 represents a tree in star format (no relation between the tips of a phylogeny), and a λ=1 would represent the original tree. The test consists in identifying by a Maximum Likelihood procedure the tree transformation – by changing the λ value – which best correlates with a Brownian Motion (BM) evolutionary model of the observed trait. So the result is generally between 0 (absence of phylogenetic signal) and 1 (a BM evolutionary model can explain 100% of the trait values). The Blombergs’s K is calculated as the variance of the observed trait divided by the expected variance under a specific evolutionary model of trait evolution (here we assume a BM). A value of K equal to 1 would characterize that the observed trait fits the BM expectation.

For all phylogenetic signal analysis we considered the phylogenetic distance as the pairwise branch length distance (although, as said before, Abouheif’s Cmean consider only topological distances). The significance of both the “statistic” and “model-based” tests were calculated by comparing the observed results against a null distribution. To generate the null distribution for each respective phylogenetic signal metric, we repeated 999 times a procedure in which we permuted the trait values and calculated the correspondent phylogenetic signal metric.

In addition, we applied a phylogenetic signal-representation (PSR) curve, an eigenvector-based evaluation proposed by Diniz-Filho et al. (2012a) that allows a more specific understanding of the evolutionary process underlying the extinction risk. This analysis consists in decomposing the phylogenetic distance matrix into eigenvectors by a multivariate technic equivalent to a PCoA. Subsequently, these eigenvectors (of length = nº of species – 1) are commutatively regressed against the trait values. The coefficient of determination (R2) generated in each regression is then compared against the cumulative eigenvalues correspondent to each eigenvector. If the trait follows a BM model of evolution, a 45° line is expected, a deviation above the curve indicates a trait evolution faster than BM, by the other hand, a curve beyond such line represents a trait evolution slower than a BM or a null model. Note that we are not considering that extinction risk ‘evolve’, like body size for example, however we assume that extinction risk is connected to other traits that evolve, which finally would lead to an evolutionary pattern. To verify the significance of the test, we calculated the area that is created by the observed curve and a BM expectation (the 45º line) 83

and after compared it against a null distribution. To make the null distribution, we permuted the trait values, then generated a new curve and calculate its correspondent area, the procedure was repeated 99 times.

The frequency distributions of proportion of threatened species per genus were almost binary for both the Iberian Peninsula and for Eastern Andalusia (see figure 3 and 6 of the supplementary material). For this reason, we also applied a binary metric of phylogenetic signal. We chose the D metric proposed by Fritz & Purvis (2010) which was primarily designed to assess the phylogenetic aggregation of extinction risk. The D is calculated given:

∑ �� − ��(∑ ��) � = ��(∑ ��) − ��(∑ ��)

Where d is the trait difference between sister clades, dBM is the value expected under a BM evolutionary model and drandom the value expected by randomizing the traits. For generating both expected values we permuted the trait values 1000 times. A value of D of 0 would indicate a risk aggregation following a BM model (below 0, extremely clumped) and a value of 1 a random association (over 1, overdispersal). Since our data is not completely binary, we applied thresholds to calculate the D and its significance. To do so, instead of choosing one value we applied a continuous threshold from 0 to 1, in 100 steps of 0.01.

(B) Phylogenetic traits

For all genera, we generated the following phylogenetic traits: species richness, taxon age, diversification rate and evolutionary distinctiveness. We tried to follow the same metrics used by Davies et al. (2011) in order to keep our results comparable. The clade species richness was calculated as the total number of species in each genus. Taxon age was calculated as the most recent common ancestral node distance to the root, which here represents the antiquity of the clade. Note that this metric can still be susceptive to the problem of a biased tree prune as the phylogeny does not include fossil data (Gaston & Blackburn, 1997; Johnson et al., 2002; see Table S1). Diversification rate is set as the amount of species in the taxa divided by its age, in order that an old taxon with few species has a low diversification rate, while a new taxon with many species has a higher diversification rate:

�� (�) �� = �

Where s is the species richness, and t the taxon age. Note that the DR (Diversification Rate) represents the balance between extinctions and speciation, we cannot disentangle both metrics as the phylogeny does not includes fossil data. Also, the DR represents the balance for the region and if a clade includes more species with a distribution outside the study area, it would not be accounted here. This is because we want a diversification rate estimation specific for the region.

Evolutionary distinctiveness is the amount of unique evolutionary history that the taxa represent. For each tip of the phylogenetic tree a value of evolutionary distinctiveness was calculated following the proposition of Isaac et al. (2007). The calculation is done by the following formula:

�� = �� + ∑ ��

Where BLtip is the tip branch length and BLi represents the ancestral branch length i of the target tip pondered by di, the number of descendants tips from the branch i. Prior to the analysis, we log transformed the phylogenetic traits so they could better fit a normal distribution and to avoid the overweight of clades with higher values on the analysis.

(B) Statistical analysis

To understand how the predictor variables correlate with each other, we calculated the pairwise Pearson correlation between all phylogenetic traits. To avoid problems of collinearity in our statistical models, we reduced our initial phylogenetic traits into orthogonal variables (eigenvectors) by applying a Principal Component Analysis (PCA). We selected the amount of eigenvectors that together would represent at least 80% of the phylogenetic traits variation. To verify the connection between the phylogenetic traits generated (see above) and the proportion of species at extinction risk by genus, we applied a Multiple General Linear Model (GLM) using the PCA axis as predictors, assuming a binomial error. Note that we assumed

a binomial error following the distribution of our response variables (see figure 3 and 6 of the supplementary material).

(B) Phylogenetic diversity loss scenario

To assess the amount of evolutionary loss (here considered as phylogenetic diversity loss) in both regions, we applied an adapted approach from the original proposition of Batista et al. (2013) to generate extinction scenarios. For this, they suggest an algorithm in which the tips are removed from the phylogeny at each step following their risk probability until no species is left. This would generate a curve of expected extinction, what, at the end, would be compared against a null expectation. The algorithm is described the 4 steps bellow:

1. The genus with the highest proportion of threatened species is selected. If more than one genus have the maximum value, the choice is randomly made between them;

2. The selected genus in the step 1 is pruned from the phylogenetic tree;

3. The phylogenetic diversity (PD) is calculated for the new phylogenetic tree. The PD is calculated as the total sum of branch lengths;

4. The process is repeated until no genus is left.

Because of the uncertainty generated in the first step (when species have to be picked at random), the complete process was repeated 100 times. We generate an average curve of expected extinction and calculate the area under the curve (AUC). Once we have the observed value, we compared it against a null distribution. The null distribution was generated by first permuting the tip values, after applying the described algorithm and then calculating the AUC. The procedure was repeated 99 times. If the observed value was lower than the null distribution a disproportional loss of evolutionary history is expected, while a value equal to the null distribution would represent a random loss of evolutionary history. In case the value is higher than the null distribution, then a lower loss of evolutionary history is expected.

(B) Data availability and analysis reproducibility

All data used in this work is available as supplementary material:

• Two text files, containing the species lists of both regions and their correspondent classification code following the IUCN regional criteria (Appendix S3 and S4). • One nexus file, including the phylogenetic hypothesis used in the work (Appendix S5). All analysis made for this manuscript were done on the free software R (RCore, 2015). To keep our analysis reproducible, we described all the code for the analysis in a pdf file (generated from an R markdown code) available as supporting information (Appendix S1).

(A) RESULTS

(B) Phylogenetic signal in extinction risk

None of the “statistic” or “model-based” metrics offered evidence for the presence of phylogenetic signal in the Spanish Iberian Peninsula flora (Table 1). Neither the Moran’s I correlogram showed any phylogenetic aggregation at any distance class (see figure 11 of the

Appendix S1). Even though the Abouheif’s Cmean was statistically significant (p = 0.034), the observed effect was too small to be considered a relevant phylogenetic aggregation (C = 0.041). The PSR curve found for the Spanish Iberian Peninsula flora showed a pattern that indicates a trait evolution slower than expected by Brownian Motion (p = 0.02), being slightly different from the effect size under the null expectations (Fig. 2A). The D metric results for binary traits showed some significant phylogenetic signal results when applying a threshold at lower values of the proportion of threatened species per genus. This indicates that a phylogenetic signal appeared in the Iberian Peninsula flora when all genus that had at least one species at risk of extinction are considered threatened. However, the significant values of D were too close to 1, indicating merely a very small signal (Fig. 3A).

For the Eastern Andalusia region, the “statistic” metrics of phylogenetic signal showed no autocorrelation (Table 1). Even the Moran’s I correlogram revealed, virtually, no autocorrelation (see figure 12 of the Appendix S1). By the other hand, both “model-based” metrics exhibited significant values, with the Pagel’s λ showing a slight small signal (λ =

0.209; p = 0.0041) and almost no effect from the Blomberg’s K (K = 0.0212; p = 0.016). The PSR curve for Eastern Andalusia presented a pattern indicating an evolution model similar to the one found for the Iberian Peninsula (Fig. 2B), i.e. slower than the expected under a Brownian Motion trait evolution (p = 0.01). The D metric results showed that a phylogenetic signal appears only when applying the threshold at the higher levels of proportion of threatened species (i.e. considering threatened only the genus with more than 60% of threatened species). However, the results of D have not gone below 0.9 (Fig. 3B), representing a small phylogenetic signal.

(B) Phylogenetic traits effects on the extinction risk

For both the Spanish Iberian Peninsula and Eastern Andalusia flora, genus species richness was only correlated with diversification rate (Iberian Peninsula: r = 0.67; Eastern Andalusia: r = 0.54). Taxon age was highly correlated with evolutionary distinctiveness (Iberian Peninsula: r = 0.91; Eastern Andalusia: r = 0.91) and negatively correlated with diversification rate (Iberian Peninsula: r = -0.50; Eastern Andalusia: r = -0.53). Accordingly, genera with higher diversification rates were species richer, but presented less evolutionary distinctiveness and were younger. Older genera were more evolutionary distinct and revealed lower rates of diversifications, with no relationship with species richness (see Table 5 and 7 of the Appendix S1 for the complete correlation matrices). With the PCA technique, we could reduce these correlated phylogenetic traits of both regions to two independents eigenvectors that together explained about 84% of the total plant traits variation from the Spanish Iberian Peninsula and to 82% from the Eastern Andalusia (Table 2).

The two first PCA axis were included as predictor variables in the statistical analysis. The GLM results for the Iberian Peninsula showed no significant values (Table 3). Therefore, no effect of evolutionary history was detected for this region. By the other hand, for the Eastern Andalusia, there is a significant negative effect of the first eigenvector on the response variable (Table 3). This eigenvector is negatively correlated with taxon age and evolutionary distinctiveness, with a small positive contribution of diversification rate (Table 2). Thus, based on the results, in the Eastern Andalusia flora the proportion of threatened

species per genus: (I) increases towards older and evolutionary distinct clades; and with an much smaller effect (II) decreases towards clades with low diversification rates.

(B) Disproportional loss of evolutionary history

The extinction scenario generated for the Iberian Peninsula plants presented an expected loss of phylogenetic diversity not different from what was estimated for the null scenarios (p = 0.075; Fig. 4A). Conversely, the extinction scenario of Eastern Andalusia flora indicates a higher loss of phylogenetic diversity when compared to the null expectation (p = 0.005; Fig. 4B). As a consequence, if the current distribution of extinction risk in the Eastern Andalusia flora determines the future extinctions, the plants of the region will probably face a disproportional loss of evolutionary history. When the extinction scenario removes around 28% (0.72 in the x-axis of the Fig. 4B) of the plant genera in Eastern Andalusia model, the expected phylogenetic diversity is far behind than the null expectation. After this point the curve starts to be more similar with the null expectation. This happens because 28.82% is the amount of genera that have at least one species threatened.

(A) DISCUSSIONS

(B) Phylogenetic clustering in extinction risk in Eastern Andalusia but not in the whole Spanish Iberian Peninsula

The well-known relationship between extinction vulnerability and phylogenetic conserved traits (e.g. body size, dispersion capability, offspring size, temperature tolerance), in animals, is likely the underlying factor of a phylogenetic signal in extinction risk (Fritz & Purvis, 2010). That is to say that any trait that is entirely or partially related to the evolutionary history of a group, and that affects direct or indirectly its species persistence, would finally drive the extinction risk to follow the same phylogenetic pattern as its own. Naturally, the multivariate causes of extinction risk and the inaccuracy associated to both the measurement of the real risk and the ideal phylogenetic relationships, would lead to noises on the observed phylogenetic signal.

Evidences of a phylogenetic signal in plants are contradictory. The absence of a phylogenetic signal was indicated for the extinction risk in mangroves species (Daru et al., 2013) and the Great Britain and South Africa cape’s flora (Davies et al., 2011). Davies et al. (2011) argued that the only trait identified to correlate with extinction risk, so far, was geographical range size, which does not present a phylogenetic signal. However, for the African’s Eastern Arc Mountain flora, Yessoufou et al. (2012) showed that a phylogenetic signal exists in both the generic and specific level. Moreover, in a more recent study at the Brazilian Atlantic Forest, Leão et al. (2014) found that vegetation type and growth form were correlated with extinction risk leading to the presence of a phylogenetic signal. Beyond all the research differences that could imply in such divergent conclusions, a complication factor is the different choices for measuring the phylogenetic signal that can have a clear influence on the results (Diniz-Filho et al., 2012b; Münkemüller et al., 2012).

Here we benefit from a general methodological view to provide more comprehensive conclusions. The absence of phylogenetic signal in the extinction risk of the Iberian Peninsula flora could indicate that the current risk is not associated to any phylogenetically clustered trait. So, factors that are independent of evolutionary history shall be the main drivers of current extinction risk in the region. Therefore, for Eastern Andalusia, the results are not conclusive. The Pagel’s λ showed a small signal on such risk. Following the revision of Münkemüller et al. (2012), small effects of phylogenetic signal were better detected by this metric, what could be the explanation for the observed results. Furthermore, the PSR and D metric results supports the presence of a small signal. So, considering that the signal also suffers the detection noise cited before, it is reasonable to cogitate that for Eastern Andalusia the presence of a phylogenetic signal is very likely to occur. This indicates that there is possibly an effect from one or more phylogenetic conserved traits on the extinction risk of the region.

(B) Biogeography intermediates the phylogenetic trait effects on the extinction risk

Our results seem to be inconsistent with the hypothesis linking the higher rates of peripatric evolution in plants to a greater extinction risk in highly diversifying and younger clades. Our finds were also not congruent between the both scales studies. While in the Iberian Peninsula

no pattern was found, the results in Eastern Andalusia showed coherence with the type 1 pattern effect of evolutionary history. The broader region (Iberian Peninsula) does not share the same evolutionary effects of the restricted one (Eastern Andalusia). The answer for such different results can be found by comparing the species statuses differences between both regions. By doing so, we noticed that 136 species that are considered threatened in Eastern Andalusia are not threatened in the Iberian Peninsula. Most of these species reach Eastern Andalusia as the southern edge of their distribution range, where they found a refuge after the climate became dryer during the Pliocene-Pleistocene transition (boreo-alpine distribution) (Médail & Diadema, 2009). Species that shows a bored-alpine distribution are necessarily older than narrower endemic taxa that are also threaten in Andalusia. In sum, the pattern disappears in the Spanish Iberian Peninsula because the boreo-alpine taxa that are threatened in Andalusia are out of risk at the scale of the Iberian Peninsula. This is also corroborated as taxon age and evolutionary distinctiveness are highly correlated, meaning that the effect must been happening for a long period.

The idea of environmental conditions changes mediating the effects of evolutionary history in Andalusia plants communities can be traced back to Herrera (1992). He found that older lineages evolved under a Pre-Mediterranean climate, and still holds trait syndromes adapted to such conditions. The occurrence of these older lineages in the region represents what he called ecological phantoms, i.e. species with traits adapted to pre-existing conditions. His ideas were them confirmed for all Mediterranean flora (Verdú et al., 2003). Furthermore, Herrera (1992) empirically noticed that these old lineages were slowly going extinct while new lineages more spread, however no statistical support was given. Our results, gives this support to Herrera (1992) hypothesis of threatened ecological phantoms in Andalusia.

The comprehension of such reasons brings new insights into the possible causes leading to an effect of evolutionary history on current extinctions, one that did not receive much attention until now (see Table S1), the biogeographic history effect. It seems that the climate instability of the region is a key factor in determining the persistence of boreo-alpine species in Eastern Andalusia. Indeed, paleoclimate are well known to influence species current distribution and consequently its extinction risk, but the idea was hardly considered in the context of the problem faced here (Svenning et al., 2015). Yet, we cannot assume that climate change alone is the only explanation for the observed results. We speculate that a 91

synergy of causes is more prone to be driving the pattern, mediated by the biogeographic history.

If older species are more specialized, we can assume that they are better adapted to specific environments and for the exposed biological factors in the region, making these species, in general, better competitors under optimum available conditions when compared to generalists. However, specialized species are also more vulnerable to environmental changes, mainly if this transformation is faster (time and specialization hypothesis in Table S1). Therefore, being more specialized (or older, if we assume the relation) can be better or worse for species long-term persistence depending on the biogeographic history of the region. In environmentally stable regions, specialization would provide advantage in survivorship, while in unstable regions being specialized can be detrimental. Considering the proposition of Davies et al. (2010), under an environmental stable scenario, peripatric speciation would lead to small geographic ranges. These small-range species would find more competition by older and well established species (specialists) in regions with long-term stable conditions. Additionally, speciation would be more prone to take place in not occupied peripheral environmental conditions, which would probably keep harder a range expansion. Whereas, in an environmental unstable region, newer and generalist species, independent of the starting range size, would have more opportunities for geographical range expansion (we provide a theoretical framework in Fig. 5).

At large scales, different biogeographical histories would confound the results, probably causing more complex patterns that would seem to disappear in analysis that do not consider biogeographic aspects. Additionally, human impacts are more likely to erase such patterns outcome at large scales, as they are an immediate cause of extinction. This seems to be a possible explanation for the absence of pattern in the Iberian Peninsula and perhaps in other studies that faced the same question (e.g. Schwartz, 1993; Vamosi & Wilson, 2008; Arregoitia et al., 2013). Therefore, such theoretical framework proposed here is a pos-hoc hypothesis, which still needs to be formally tested to be confirmed.

(B) The expected impact on evolutionary history

Following the predictions, the absence of a phylogenetic signal (no risk aggregation) and no higher risk towards older or unique clades in the Iberian Peninsula did not caused a greater loss of evolutionary history from what would be expected under a null model. For Eastern Andalusia, however, the results of the extinction model indicates a disproportional loss of phylogenetic diversity, due to (or because) the presence of a phylogenetic signal and a type 1 pattern. Older clades are also more phylogenetic distinct, and therefore, they represents more unique amounts of evolutionary history. Losing such clades costs a lot in term of evolution. If we consider that the risk is clustered in these clades, as pointed out for Eastern Andalusia, we can affirm that future extinctions will be more liable to remove more evolutionary history than if there was no evolutionary pattern in extinction risk at all.

The possible imminence of such a loss of phylogenetic diversity in the region does not really indicate a conservation problem. First, because the local extinctions of these specific boreo-alpine flora does not represent a global extinction of the species; and, second, it is part of a long-term natural process that is happening in the region, not a drastic rapidly change in the composition, as caused by humans activities. However, assuming the phylogenetic conservatism of traits in plants (Prinzing, 2001), we can also expect for the region a decrease in its functional diversity. Also, we can suppose that as older species (and maybe more specialist) will be lost, other species that exclusively interacts with them shall disappear as part of a coextinction process (Vieira & Almeida‐Neto, 2015).

(A) CONCLUSIONS

Our research suggests that in Eastern Andalusia older lineages carrying more unique phylogenies are more vulnerable to extinction. This is probably driven by the presence of Pre-Mediterranean species that reach the region with its southern most distribution. The analysis also indicates the presence of a phylogenetic signal in extinction risk for the Eastern Andalusia region. The presence of such risk aggregation towards old and unique clades indicates a possible disproportional loss of evolutionary history in the region. These patterns were not obtained when we upscaled the analysis to the Iberian Peninsula. Our proposed framework suggests that the biogeographical history underlies these results and we suggest 93

that it should be considered in future works examining the evolutionary role in extinctions. Moreover, the results presented here serve also as an example of the effects of past climate changes on extinctions and species geographical range dynamic, providing some expectations for future events. Finally, as knowledge shortfalls are being overcame, more possibilities appears in integrating evolutionary history and also biogeographic aspects not only to fully understand the current extinction patterns, but to understand the broad extinction process itself.

(A) ACKNOWLEDGEMENTS

We thanks J.M.C. Ortega, L. Pataro, A. Briega, F. Villalobos, R. Dobrovolski, S. Gouveia for insightful comments on the ideas of the manuscript. We also thanks E. Sideri for comments on grammar issues. BV thanks the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for a doctoral scholarship.

(A) REFERENCES

Abouheif, E. (1999) A method for testing the assumption of phylogenetic independence in comparative data. Evolutionary Ecology Research, 1, 895-909.

Alroy, J. (2015) Current extinction rates of reptiles and amphibians. Proceedings of the National Academy of Sciences, 112, 13003-13008.

Arregoitia, L.D.V., Blomberg, S.P. & Fisher, D.O. (2013) Phylogenetic correlates of extinction risk in mammals: species in older lineages are not at greater risk. Proceedings of the Royal Society of London B: Biological Sciences, 280, 20131092.

Birand, A., Vose, A. & Gavrilets, S. (2012) Patterns of species ranges, speciation, and extinction. The American Naturalist, 179, 1-21.

Blanca, G., Cabezudo, B., Cueto, M., Fernández López, C. & Morales Torres, C. (2009) Flora vascular de Andalucía oriental, 4 vols. Consejería de Medio Ambiente, Junta de Andalucía, Sevilla.

Blomberg, S.P., Garland Jr, T., Ives, A.R. & Crespi, B. (2003) Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution, 57, 717-745.

Brusatte, S.L., Butler, R.J., Barrett, P.M., Carrano, M.T., Evans, D.C., Lloyd, G.T., Mannion, P.D., Norell, M.A., Peppe, D.J. & Upchurch, P. (2015) The extinction of the dinosaurs. Biological Reviews, 90, 628-642.

Daru, B.H., Yessoufou, K., Mankga, L.T. & Davies, T.J. (2013) A global trend towards the loss of evolutionarily unique species in mangrove ecosystems. PloS one, 8, e66686.

Davies, T.J. (2015) Losing history: how extinctions prune features from the tree of life. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 370, 20140006.

Davies, T.J., Smith, G.F., Bellstedt, D.U., Boatwright, J.S., Bytebier, B., Cowling, R.M., Forest, F., Harmon, L.J., Muasya, A.M. & Schrire, B.D. (2011) Extinction risk and diversification are linked in a plant biodiversity hotspot. PLoS Biology, 9, e1000620.

Diniz-Filho, J.A.F., Rangel, T.F., Santos, T. & Mauricio Bini, L. (2012a) Exploring patterns of interspecific variation in quantitative traits using sequential phylogenetic eigenvector regressions. Evolution, 66, 1079-1090.

Diniz-Filho, J.A.F., Santos, T., Rangel, T.F. & Bini, L.M. (2012b) A comparison of metrics for estimating phylogenetic signal under alternative evolutionary models. Genetics and molecular biology, 35, 673-679.

Fahrig, L. (1997) Relative effects of habitat loss and fragmentation on population extinction. The Journal of Wildlife Management, 61, 603-610.

Finnegan, S., Heim, N.A., Peters, S.E. & Fischer, W.W. (2012) Climate change and the selective signature of the Late Ordovician mass extinction. Proceedings of the National Academy of Sciences, 109, 6829-6834.

Fritz, S.A. & Purvis, A. (2010) Selectivity in mammalian extinction risk and threat types: a new measure of phylogenetic signal strength in binary traits. Conservation Biology, 24, 1042-1051.

Gaston, K.J. & Blackburn, T.M. (1997) Evolutionary age and risk of extinction in the global avifauna. Evolutionary Ecology, 11, 557-565.

Gittleman, J.L. & Kot, M. (1990) Adaptation: statistics and a null model for estimating phylogenetic effects. Systematic Biology, 39, 227-241.

Herrera, C.M. (1992) Historical effects and sorting processes as explanations for contemporary ecological patterns: character syndromes in Mediterranean woody plants. American Naturalist, 140, 421-446.

Hinchliff, C.E. & Smith, S.A. (2014) Some Limitations of Public Sequence Data for Phylogenetic Inference (in Plants). PLoS ONE, 9, e98986.

Hodgson, J. (1986) Commonness and rarity in plants with special reference to the Sheffield flora Part III: taxonomic and evolutionary aspects. Biological conservation, 36, 275- 296.

Huang, S., Davies, T.J. & Gittleman, J.L. (2011) How global extinctions impact regional biodiversity in mammals. Biology letters, 8, 222-225.

Isaac, N.J., Turvey, S.T., Collen, B., Waterman, C. & Baillie, J.E. (2007) Mammals on the EDGE: conservation priorities based on threat and phylogeny. PLoS One, 2, e296.

IUCN (2001) IUCN red list categories and criteria: version 3.1, Gland, Switzerland.

IUCN (2010) Guidelines for Application of IUCN Red List Criteria at Regional Levels (V. 4.0), Gland, Switzerland.

IUCN (2015) IUCN red list of threatened species. Version 2015.2. Acessed on 12 November 2015. In, http://www.iucnredlist.org. .

Jaeger, R.G. (1970) Potential extinction through competition between two species of terrestrial salamanders. Evolution, 24, 632-642.

Johnson, C.N., Delean, S. & Balmford, A. (2002) Phylogeny and the selectivity of extinction in Australian marsupials. Animal Conservation, 5, 135-142.

Leão, T.C., Fonseca, C.R., Peres, C.A. & Tabarelli, M. (2014) Predicting extinction risk of Brazilian Atlantic Forest angiosperms. Conservation Biology, 28, 1349-1359.

McDowall, R. (1969) Extinction and endemism in New Zealand land birds. Tuatara, 17, 1- 12.

McKinney, M.L. (1997) How do rare species avoid extinction? A paleontological view. The biology of rarity: causes and consequences of rare-common differences (ed. by W.E. Kunin and K. Gaston). Springer Science & Business Media.

McKinney, M.L. & Lockwood, J.L. (1999) Biotic homogenization: a few winners replacing many losers in the next mass extinction. Trends in ecology & evolution, 14, 450-453.

Médail, F. & Diadema, K. (2009) Glacial refugia influence plant diversity patterns in the Mediterranean Basin. Journal of Biogeography, 36, 1333-1345.

Meijaard, E., Sheil, D., Marshall, A.J. & Nasi, R. (2008) Phylogenetic age is positively correlated with sensitivity to timber harvest in Bornean mammals. Biotropica, 40, 76- 85.

Molina‐Venegas, R. & Roquet, C. (2014) Directional biases in phylogenetic structure quantification: a Mediterranean case study. Ecography, 37, 572-580.

Moran, P.A. (1950) Notes on continuous stochastic phenomena. Biometrika, 37, 17-23.

Münkemüller, T., Lavergne, S., Bzeznik, B., Dray, S., Jombart, T., Schiffers, K. & Thuiller, W. (2012) How to measure and test phylogenetic signal. Methods in Ecology and Evolution, 3, 743-756.

Pagel, M. (1999) Inferring the historical patterns of biological evolution. Nature, 401, 877- 884.

Pavoine, S., Ollier, S., Pontier, D. & Chessel, D. (2008) Testing for phylogenetic signal in phenotypic traits: new matrices of phylogenetic proximities. Theoretical population biology, 73, 79-91.

Prinzing, A. (2001) The niche of higher plants: evidence for phylogenetic conservatism. Proceedings of the Royal Society of London B: Biological Sciences, 268, 2383-2389.

Purvis, A. (2008) Phylogenetic approaches to the study of extinction. Annual Review of Ecology, Evolution, and Systematics, 39, 301-319.

Purvis, A., Agapow, P.-M., Gittleman, J.L. & Mace, G.M. (2000) Nonrandom extinction and the loss of evolutionary history. Science, 288, 328-330.

Quental, T.B. & Marshall, C.R. (2013) How the Red Queen drives terrestrial mammals to extinction. Science, 341, 290-292.

Raup, D.M. (1994) The role of extinction in evolution. Proceedings of the National Academy of Sciences, 91, 6758-6763.

Raup, D.M. & Sepkoski, J.J. (1982) Mass extinctions in the marine fossil record. Science, 215, 1501-1503.

RCore, T. (2015) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. In. ISBN 3-900051-07-0, URL http://www.R-project.org

Safi, K. & Pettorelli, N. (2010) Phylogenetic, spatial and environmental components of extinction risk in carnivores. Global Ecology and Biogeography, 19, 352-362.

Saiz, J.C.M., Lozano, F.D., Gómez, M.M. & Baudet, Á.B. (2015) Application of the Red List Index for conservation assessment of Spanish vascular plants. Conservation Biology, 29, 910-919.

Schwartz, M.W. (1993) The search for pattern among rare plants: are primitive species more likely to be rare? Biological Conservation, 64, 121-127.

Simberloff, D. (1986) The proximate causes of extinction. Patterns and processes in the history of life, pp. 259-276. Springer.

Sinclair, A.R.E., Pech, R.P., Dickman, C.R., Hik, D., Mahon, P. & Newsome, A.E. (1998) Predicting Effects of Predation on Conservation of Endangered Prey. Conservation Biology, 12, 564-575.

Smith, K.G., Lips, K.R. & Chase, J.M. (2009) Selecting for extinction: nonrandom disease‐ associated extinction homogenizes amphibian biotas. Ecology Letters, 12, 1069-1078.

Svenning, J.-C., Eiserhardt, W.L., Normand, S., Ordonez, A. & Sandel, B. (2015) The Influence of Paleoclimate on Present-Day Patterns in Biodiversity and Ecosystems. Annual Review of Ecology, Evolution, and Systematics, 46

TPL (2013) The Plant List. Version 1.1. Acessed on 12 November 2015. In, http://www.theplantlist.org/.

Vamosi, J.C. & Wilson, J.R. (2008) Nonrandom extinction leads to elevated loss of angiosperm evolutionary history. Ecology letters, 11, 1047-1053.

Verdú, M., Dávila, P., García‐Fayos, P., Flores‐Hernández, N. & Valiente‐Banuet, A. (2003) Convergent traits of mediterranean woody plants belong to pre‐mediterranean lineages. Biological Journal of the Linnean Society, 78, 415-427.

Veron, S., Davies, T.J., Cadotte, M.W., Clergeau, P. & Pavoine, S. (2015) Predicting loss of evolutionary history: Where are we? Biological Reviews,

Vieira, M.C. & Almeida‐Neto, M. (2015) A simple stochastic model for complex coextinctions in mutualistic networks: robustness decreases with connectance. Ecology letters, 18, 144-152.

Wang, S., Chen, A., Fang, J. & Pacala, S.W. (2013) Why abundant tropical tree species are phylogenetically old. Proceedings of the National Academy of Sciences, 110, 16039- 16043.

Yessoufou, K., Daru, B.H. & Davies, T.J. (2012) Phylogenetic patterns of extinction risk in the Eastern Arc ecosystems, an African biodiversity hotspot. PLoS One, 7, e47082.

100

Table 1. Multiple results for the phylogenetic signal in the proportion of threatened species per genus in vascular plants of the Spanish Iberian Peninsula and Eastern Andalusia with their correspondent p-value.

IBERIAN PENINSULA ANDALUSIA Metric Observed p-values Observed p-values

Moran's I 0.0395 0.0500 0.0236 0.1590 Abouheif's C -0.0410 0.0340 0.0250 0.1740 Pagel's λ 0.0906 0.1145 0.2099 0.0041 Blomberg's Κ 0.0174 0.0430 0.0212 0.0160

101

Table 2. Principal Component Analysis of the plants phylogenetic traits of Spanish Iberian Peninsula and Eastern Andalusia. The table shows the loadings (i.e. the correlation between the original variable and the correspondent eigenvector) of each variable in relationship to the chosen orthogonal eigenvectors. Percentage values indicate the amount of original variation captured by the axis.

IBERIAN PENINSULA ANDALUSIA Phylogenetic trait PCA1 (45%) PCA2 (39%) PCA1 (47%) PCA2 (35%) Species richness 0.19 0.97 -0.02 0.98 Taxon age -0.75 0.16 -0.77 0 Evolutionary distinctiveness -0.62 0.13 -0.63 0.01 Diversification rate 0.14 0.15 0.13 0.19

102

Table 3. Multiple Generalized Linear Models including the two first PCA axis generated from the phylogenetic traits for the vascular plants of Spanish Iberian Peninsula and Eastern Andalusia.

Region Explanatory variable Coefficient Z p value

Iberian Peninsula PCA1 -0.072 -1.151 0.25

PCA2 -0.031 -0.417 0.676

Eastern Andalusia PCA1 -0.149 -2.027 0.042

PCA2 -0.133 -1.220 0.225

103

Figure 1. Map of the studied areas. In green the Spanish Iberian Peninsula region, and in red Eastern Andalusia, a Mediterranean sub region comprised by the former.

104

Figure 2. Phylogenetic Signal Representation curve (PSR) for the Spanish Iberian Peninsula (A) and the Eastern Andalusia (B) flora. Black dots represent the observed values. Gray lines represent the null expectation based on the 99 permutation results. The straight black line represents the expectation of a Brownian Motion (BM) model of evolution. Both curves are different from the null expectation, Spanish Iberian Peninsula (p = 0.02) and Eastern Andalusia (p = 0.01). They fit a model slower than the expected by a BM.

105

Figure 3. D value calculated to check the phylogenetic signal of proportion of threatened species for the Spanish Iberian Peninsula (A) and Eastern Andalusia (B) vascular plant genera along a gradient of 100 thresholds. Results with p-values below 0.05 are colored black.

106

Figure 4. Extinction loss model curve for the Spanish Iberian Peninsula (A) and Eastern Andalusia (B) plant genus. The graph shows the expected phylogenetic diversity along a process of extinction in which the most threatened genus is eliminated per time until no genus is left (genus proportion = 0). Black line represent the mean observed values and gray lines the values from the null expectation.

107

Figure 5. Theoretical scheme integrating biogeographical aspects into current relationships between evolutionary history and current extinction risks. (A) Initial closed environmental space, with its correspondent available conditions on the geographic space. Continuous circles represent the 100% conditions available in geography, while the dashed circles 50% most abundant conditions. Species Grinnellian niches are represented by blue forms. (B) Geographically available conditions remain the same as initial. New species (orange striped forms) find more competition and tend to appear in more peripheral conditions. (C) Environmental condition changes in geographic space. New species find less competition and unoccupied conditions. (D) The effects of the different scenarios on the tree of life extinction risk. The scenario in B would lead to younger species to be more threatened while the scenario in C would lead to the inverse pattern.

108

SUPPORTING INFORMATION*

*Due the size of the supporting information, we made then all available at the following link: http://1drv.ms/1mwWLz3

Appendix S1. Analyses code description. Appendix S2. Description of the studied areas. Appendix S3. Plant red list of the Iberian Peninsula. Appendix S4. Plant red list of the Oriental Andalucía. Appendix S5. Phylogenetic tree for the Iberian Peninsula flora. Table S1. Hypotheses list of phylogenetic patterns of extinction risk.

109

Capítulo 4: letsR: a new R package for data handling and analysis in macroecology

Bruno Vilela e Fabricio Villalobos

Artigo publicado na revista Methods in Ecology and Evolution em maio de 2015 DOI: 10.1111/2041-210X.12401

*Devido ao tamanho do Material Suplementar deste artigo (Supporting information), o mesmo deverá ser acessado online, seguindo o DOI do artigo ou o seguinte link: http://goo.gl/W0YnxQ

110

Summary

1. The current availability of large ecological data sets and the computational capacity to handle them has fostered the testing and development of theory at broad spatial and temporal scales. Macroecology has particularly benefited from this era of big data but tools are still required to help transforming this data into information and knowledge. 2. Here, we present ‘letsR’, a package for the R statistical computing environment, designed to handle and analyze macroecological data such as species’ geographic distributions (polygons in shapefile format and point occurrences) and environmental variables (in raster format). The package also includes functions to obtain data on species’ habitat use, description year, and current as well as temporal trends in conservation status as provided by the IUCN RedList online database. 3. ‘letsR’ main functionalities are based on presence-absence matrices that can be created with the package’s functions and from which other functions can be applied to generate, for example, species richness rasters, geographical midpoints of species, and species- and site- based attributes. 4. We exemplify the package’s functionality by describing and evaluating the geographic pattern of species’ description year in tailless amphibians. All data preparation and most analyses were made using the ‘letsR’ functions. Our example illustrates the package’s capability for conducting macroecological analyses under a single computer platform, potentially helping researchers to save time and effort in this endeavor.

Key-words: biodiversity gradients, geographic distribution, presence-absence matrix, spatial analysis, species richness

111

Introduction

The rise of “big data” is leading to transformations in the way research is conducted in different knowledge areas and ecology is not an exception (Hampton et al. 2013). Such amount and variety of data and the computational capacity to process it has turned ecology into a data-intensive science (Michener & Jones 2012), enabling theory testing and development at broader spatial and temporal scales and with more resolution than any time before. For instance, the macroecological research program with its focus on the statistical regularities emerging from studying ecological systems at large scales (Marquet 2009) has particularly benefited from this big data revolution. Indeed, macroecology has steadily grown and become central in ecological research (Beck et al. 2012) contributing answers to unsolved questions in ecology and evolution, some of which were stated centuries ago (Hawkins 2001; Gaston, Chown & Evans 2008). This current era of big data in ecology also brings big challenges. These challenges are mainly related to the manipulation of large data sets necessary to transform such data into information and knowledge (Schadt et al. 2010; Michener & Jones 2012). Converting large data sets into information and knowledge demands different steps including data gathering, organization, preparation, analysis, and presentation (modified from Liew 2007). Moreover, without proper tools, conducting these steps may be exceedingly time consuming and easily subject to human errors. Thus, current efforts are being conducted to develop tools that can help in this endeavor (Rangel, Diniz‐Filho & Bini 2010). In macroecology, typical data involves species’ geographical distributions in polygon format or occurrence records (e.g. spatial data made available by IUCN, BirdLife, or GBIF), species traits measurements (e.g. Jones et al. 2009), phylogenetic hypotheses on species’ evolutionary relationships (e.g. Piel et al. 2000), and spatial environmental layers (e.g. Hijmans et al. 2005). Macroecological analyses require these data to be integrated and manageable. A practical way to integrate basic macroecological information is the presenceabsence matrix (PAM) (Arita et al. 2008; Gotelli et al. 2009) that summarizes species’ geographical distributions and diversity, the two fundamental units of biogeography (Arita et al. 2008). Conventionally, in a PAM, rows represent species, columns represent sites, and

112

elements record the occurrence of a given species in a given site, either as binary (i.e. presence: 1; or absence: 0) or quantitative data (i.e. abundance or trait values) (Bell 2003; Arita et al. 2008). Such PAMs can be generated from ecological surveys of localities or from overlaying a grid of cells onto the area of study, with the latter method being standard in macroecological analyses (Gotelli et al. 2009). Moreover, further data can be integrated into the PAM as additional rows representing site descriptors such as location (e.g. geographic coordinates), environmental conditions or other descriptors. In this way, a PAM allows species- and site-based analyses either summarizing data solely by species or sites (R-mode and Q-mode analyses; Simberloff & Connor 1979) or simultaneously considering information from both species and sites (Rq-mode and Qr-mode; Arita et al. 2008). Here we present a new R package, ‘letsR’, for obtaining, handling and analyzing data for macroecological research. The main functions of this package allow the user to generate a PAM from species’ geographical distributions (polygon and point occurrence data) and merge it with species’ traits and spatial environmental layers. Other package’s functions provide tools to summarize and visualize the information contained in a PAM including transformation and direct analyzes such as the estimation of distance matrices or spatial autocorrelograms. In addition, the package contains functions to acquire species data from the IUCN RedList online database such as description year (the year in which a species was described by the species authority), current conservation status and its temporal trend, and habitat use, among others. We exemplified the package’s use, functionality and capacity to convert large amounts of data into information and knowledge by analyzing the geographic pattern of description year of tailless amphibians. letsR package

The ‘letsR’ package is written in the R language and its first version was released on CRAN (The Comprehensive R Archive Network) in May 2014 and can be used under R version 2.1 or higher (RCore 2015). We are constantly updating the package and encourage interested users to look for the latest package version on GitHub (https://github.com/macroecology/letsR). The functions available in ‘letsR’ depend on other packages: ‘raster’(Hijmans 2015), ‘XML’(Lang 2013), ‘geosphere’(Hijmans 2014),

113

‘maptools’(Bivand & Lewin-Koh 2015), ‘maps’(Becker et al. 2014), ‘sp’(Bivand et al. 2008), ‘fields’(Nychka, Furrer & Sain 2014) and ‘rgdal’(Bivand, Keitt & Rowlingson 2014). Every function of ‘letsR’ starts with the prefix ‘lets.’ to avoid conflict with other R functions (see Table 1 for the main functions and their descriptions). The package was named after the Theoretical Ecology and Synthesis Lab (LETS; Portuguese acronym) of the Federal University of Goiás for their contribution to the advancement of macroecology in Brazil and worldwide.

PresenceAbsence class

The package basic functions work mainly with a new S3 object class called ‘PresenceAbsence’. The class ‘PresenceAbsence’ is generated by the three ‘letsR’ functions that create a PAM based on a user-defined grid cell system (see Presence-Absence function in Table 1). The new object class is a list consisting of three objects: (I) a sites by species matrix indicating presence (1) or absence (0) of a given species in a given site (note that this PAM arrangement is the transpose version –with sites in the rows and species in the columns— of the conventional one described above) with the first two columns containing the coordinates corresponding to the cells’ centroids; (II) a raster containing species richness values per cell; and (III) a vector with species' names contained in the matrix. Any of these three internal objects of the ‘PresenceAbsence’ class can be obtained in the standard way for objects of ‘list’ class (i.e. using list-subsetting operators: ‘[[‘ or ‘$’). The class ‘PresenceAbsence’ is fundamental for other ‘letsR’ functions, mainly because it contains information beyond the PAM itself such as the user-defined grid cell system, including its resolution, projection, datum and extent as described by the extreme coordinates, as attributes of the raster object (the second object mentioned above). These attributes are indispensable for other analysis and they cannot be stored in a simple PAM. The ‘PresenceAbsence’ class also allows using generic functions such as ‘plot’, ‘summary’ and ‘print’, which facilitate its description and visualization. Nevertheless, the Presence- absence functions (Table 1) also include an argument allowing users to choose between getting the PAM as an object of class ‘matrix’ or ‘PresenceAbsence’. Finally, some ‘letsR’ functions allow the user to input customary R objects (e.g. ‘vector’, ‘matrix’, ‘data.frame’).

114

Memory and time consumption

Generating and manipulating a PAM can consume a large amount of Random-Access Memory (RAM), mainly when handling data of large spatial extents and high resolution. The specific amount of RAM needed will depend on the grid extent, cell resolution, and the number of species in the analysis. For example, to create a PAM for all anuran species of the world (5597) with a 1° grid cell resolution (see the example analysis below), R itself can consume up to 4.7GB of RAM (note that this number can grow exponentially with the increase in cell resolution). In this example, the whole process took around 1h30min on a laptop computer with an Intel Core i7-4500U @ 1.80GHz processor and running under Windows 8.1 of 64 bits. To help the user keep track of the analysis relative running time, namely how many runs are left to finish the analysis, some functions include the ‘count’ argument to open a separate window containing the run countdown.

Example: Spatial pattern of description year in tailless amphibians (Amphibia: Anura)

To illustrate the utilities of ‘letsR’, we applied some of its functions to examine the geographic pattern of description year of anuran species (Figure 1; the detailed R code to recreate the example is available in .Rmd and .pdf formats as supporting information). More specifically we aimed to: (1) Generate a global map of the description year for the species of the Anura order; (2) Evaluate if geographically closer species show similar description year; (3) Asses the effect of the species’ geographic range size and the maximum value of human footprint (i.e. population density, land transformation, human access, and power infrastructure) within their ranges on their description years. Instead of attempting an exhaustive exploration on this subject, our main goal was to exemplify the package’s functionalities applied to a simple macroecological question using a spatially explicit approach that had only been conducted for limited areas (Diniz‐Filho et al. 2005). We obtained data on species’ geographic distributions, in the shapefile format, by manually downloading them from the IUCN online database (IUCN 2012; http://www.iucnredlist.org/technical-documents/spatial-data; Figure 1, step 1) and

115

transformed it into a presence-absence matrix (PAM) by using a global grid of 1° resolution using the function ‘lets.presab’. It should be noted that the grid itself is created by the function with its properties (e.g. resolution and extent) being chosen by the user. Applying the function ‘plot’ on the resulting ‘PresenceAbsence’ object allows the visualization of the species richness pattern (Figure 1, step 2), whereas the function ‘summary’ compiles the key information stored in the object and prints it in the console:

Class: PresenceAbsence _ _ Number of species: 5597 Number of cells: 13594 Cells with presence: 13594 Cells without presence: 0 Species without presence: 0 Species with the largest range: Bufo bufo _ _ Grid parameters Resolution: 1, 1 (x, y) Extention: -180, 180, -90, 90 (xmin, xmax, ymin, ymax) Coord. Ref.: +proj=longlat +datum=WGS84

We gathered the description year for the 5597 species in our dataset from the IUCN online database (IUCN 2012) using the function ‘lets.iucn’ (the ‘PresenceAbsence’ object can be used directly as input). We obtained the mean description year of species within each cell by applying the function ‘lets.maplizer’ (although we used the mean, other summary statistics can be applied; e.g. median, standard deviation, etc.), which returns either a vector with the cells’ coordinates and the summarized attribute or this vector and a raster containing this information. Then, we used this raster to map the geographic pattern of description year of tailles amphibians (Figure 1, step 3). Our results indicated that Europe concentrates the oldest described species, whereas tropical regions (mainly the west region of Australia, Sub- Saharan Africa and the northern Andes) contain more recently described anuran species. This pattern can be expected considering that Europe was the cradle of most early naturalists and taxonomists that pioneered the application of Linnaeus’ binomial nomenclature to species classification.

116

Exploring the geographic pattern of description year from a species-oriented perspective also provides information on the factors affecting the trend species discovery. Therefore, we obtained the species’ distributional midpoint (i.e. the geographic range centroid) using the function ‘lets.midpoint’ and generated a geographical distance matrix between species using the function ‘lets.distmat’. This distance matrix and the description year of each species were used to calculate a Moran’s I correlogram using the function ‘lets.correl’ for 200 equidistant classes – which can be equiprobable, depending on the user’s preference (Figure 1, step 4). The results of this species-oriented analysis support the idea that geographically closer species were described in proximate years. Moreover, species that are separated by a distance of 2000km (or higher) had independent description years. We also tested if the species’ description year pattern was determined by range size or human footprint level within species’ ranges. To do so, we calculated the species range size using the function ‘lets.rangesize’, which calculates the number of cells in which each species occurs (this function can also generate other range size metrics such as the polygon area or the summed area of cells). Then, we obtained a global layer of human footprint at a resolution of 30 arc-second cell size from the NASA Socioeconomic Data and Applications Center (WCS & CIESIN 2005). To assign a value of this variable to each species, we first upscaled the variable to a 1° cell size by averaging the values within each cell and then added it to the PAM as an additional column, using the function ‘lets.addvar’. After, we extracted the maximum value of human footprint within each species’ range using the function ‘lets.summarizer’ that transfers the spatial information at the cell level to the species level. Finally, we did a multiple regression with both range size and maximum human footprint within ranges as explanatory variables and species’ description year as the response variable. A Monte Carlo simulation with 999 repetitions was used to test the model’s significance. Results of this analysis indicated that both aspects of species’ ranges influence the variation of description year among anuran species, jointly explaining nearly 30% of the variance (p < 0.001).

117

Conclusion

Here we have presented the ‘letsR’ package and illustrated its functionality for conducting macroecological analyses under a single computer platform. ‘letsR’ allows handling large data quantities commonly used in macroecology and applying different tools to process and analyze such data. For instance, to our knowledge, this is the first effort showing a global map of species’ description year for an entire order of vertebrates. In addition, this also highlights the package’s potential to help researchers in testing macroecological theories. ‘letsR’ is constantly being improved to attend new demands as the field of macroecology continuous to develop.

Acknowledgements

We thank A.S. Melo, L. Sgarbi, S. Varela, J.A.F. Diniz-Filho, L. Jardim, F. Faleiro, M.S. Lima-Ribeiro, L.C. Terribile, M.A. Rodríguez, R. Dobrovolski, and S. Gouveia for useful suggestions on the package’s code and theoretical background. We also thank T. Lucas and S. Chamberlain for detailed suggestions that greatly improved our package’s code and manuscript clarity. We are grateful to all the people of LETS (Laboratório de Ecologia Teórica e Síntese – UFG) for testing earlier versions of the package. B.V. was supported by a CAPES grant for doctoral studies and F.V. by a CNPq “Science without borders” fellowship.

118

Data accessibility

Species distribution data are available in polygon format on the IUCN Red List of Threatened Species database (http://goo.gl/UOhD7Q). The global layer of human footprint at a resolution of 30 arc-second cell size is available from the NASA Socioeconomic Data and Applications Center: http://www.ciesin.columbia.edu/repository/wildareas/data/hfp_global_geo_grid.zip.

The ‘letsR’ package is available on CRAN: http://cran.r-project.org/web/packages/letsR/index.html and on Github: https://github.com/macroecology/letsR.

The R scripts to recreate the example analysis, including the function to gather the description year data (available from the IUCN Red List of Threatened Species, http://www.iucnredlist.org/) are uploaded as supporting information.

119

References

Arita, H.T., Christen, J.A., Rodríguez, P. & Soberón, J. (2008) Species diversity and

distribution in presence‐absence matrices: mathematical relationships and biological

implications. The American Naturalist, 172, 519-532.

Beck, J., Ballesteros‐Mejia, L., Buchmann, C.M., Dengler, J., Fritz, S.A., Gruber, B., Hof,

C., Jansen, F., Knapp, S. & Kreft, H. (2012) What's on the horizon for macroecology?

Ecography, 35, 673-683.

Becker, R.A., Wilks, A.R., Brownrigg, R. & Minka, T.P. (2014) maps: Draw Geographical

Maps. R package version 2.3-9. http://CRAN.R-project.org/package=maps.

Bell, G. (2003) The interpretation of biological surveys. Proceedings of the Royal Society of

London B: Biological Sciences, 270, 2531-2542.

Bivand, R., Keitt, T. & Rowlingson, B. (2014) rgdal: Bindings for the Geospatial Data

Abstraction Library. R package version 0.9-1. http://CRAN.R-

project.org/package=rgdal.

Bivand, R. & Lewin-Koh, N. (2015) maptools: Tools for Reading and Handling Spatial

Objects. R package version 0.8-34. http://CRAN.R-project.org/package=maptools.

Bivand, R.S., Pebesma, E.J., Gómez-Rubio, V. & Pebesma, E.J. (2008) Applied spatial data

analysis with R. Springer.

Diniz‐Filho, J.A.F., Bastos, R.P., Rangel, T.F., Bini, L.M., Carvalho, P. & Silva, R.J. (2005)

Macroecological correlates and spatial patterns of anuran description dates in the

Brazilian Cerrado. Global Ecology and Biogeography, 14, 469-477.

Gaston, K.J., Chown, S.L. & Evans, K.L. (2008) Ecogeographical rules: elements of a

synthesis. Journal of Biogeography, 35, 483-500.

120

Gotelli, N.J., Anderson, M.J., Arita, H.T., Chao, A., Colwell, R.K., Connolly, S.R., Currie,

D.J., Dunn, R.R., Graves, G.R. & Green, J.L. (2009) Patterns and causes of species

richness: a general simulation model for macroecology. Ecology Letters, 12, 873-886.

Hampton, S.E., Strasser, C.A., Tewksbury, J.J., Gram, W.K., Budden, A.E., Batcheller, A.L.,

Duke, C.S. & Porter, J.H. (2013) Big data and the future of ecology. Frontiers in

Ecology and the Environment, 11, 156-162.

Hawkins, B.A. (2001) Ecology's oldest pattern? Trends in Ecology & Evolution, 16, 470.

Hijmans, R. (2015) raster: Geographic data analysis and modeling. R package version 2.3-

24. http://CRAN.R-project.org/package=raster.

Hijmans, R.J. (2014) geosphere: Spherical Trigonometry. R package version 1.3-11.

http://CRAN.R-project.org/package=geosphere.

Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G. & Jarvis, A. (2005) Very high

resolution interpolated climate surfaces for global land areas. International journal of

climatology, 25, 1965-1978.

IUCN (2012) IUCN red list of threatened species. Version 2012.2. International Union for

the Conservation of Nature Gland, Switzerland.

Jones, K.E., Bielby, J., Cardillo, M., Fritz, S.A., O'Dell, J., Orme, C.D.L., Safi, K., Sechrest,

W., Boakes, E.H. & Carbone, C. (2009) PanTHERIA: a species-level database of life

history, ecology, and geography of extant and recently extinct mammals: Ecological

Archives E090-184. Ecology, 90, 2648-2648.

Lang, D.T. (2013) XML: Tools for parsing and generating XML within R and S-Plus. R

package version 3.98-1.1. http://CRAN.R-project.org/package=XML.

Liew, A. (2007) Understanding data, information, knowledge and their inter-relationships.

Journal of Knowledge Management Practice, 8, 1-16. 121

Marquet, P.A. (2009) Macroecological perspectives on communities and ecosystems. The

Princeton guide to ecology, 386.

Michener, W.K. & Jones, M.B. (2012) Ecoinformatics: supporting ecology as a data-

intensive science. Trends in Ecology & Evolution, 27, 85-93.

Nychka, D., Furrer, R. & Sain, S. (2014) fields: Tools for spatial data. R package version 7.1.

http://CRAN.R-project.org/package=fields.

Piel, W.H., Donoghue, M., Sanderson, M. & Netherlands, L. (2000) TreeBASE: a database

of phylogenetic information. Proceedings of the 2nd International Workshop of

Species 2000.

Rangel, T.F., Diniz‐Filho, J.A.F. & Bini, L.M. (2010) SAM: a comprehensive application for

spatial analysis in macroecology. Ecography, 33, 46-50.

RCore, T. (2015) R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-

project.org.

Schadt, E.E., Linderman, M.D., Sorenson, J., Lee, L. & Nolan, G.P. (2010) Computational

solutions to large-scale data management and analysis. Nature Reviews Genetics, 11,

647-657.

Simberloff, D. & Connor, E. (1979) Q-mode and R-mode analyses of biogeographic

distributions: null hypotheses based on random colonization. Contemporary

quantitative ecology and related ecometrics, 12, 123-138.

WCS & CIESIN (2005) Last of the wild project, Version 2 (LWP-2): global human footprint

dataset (IGHP). Palisades, NY: NASA Socioeconomic Data and Applications Center

(SEDAC). See http://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-

footprint-ighp (accessed 01 september 2014). 122

Table 1. List of the ‘letsR’ package main functions and their description.

Type Function Description

Presence-absence Creates a presence-absence matrix of species’ lets.presab functions geographic ranges within a grid Creates a presence-absence matrix of species’ lets.presab.birds geographic ranges within a grid for the Birdlife spatial data Creates a presence-absence matrix based on lets.presab.points species’ point occurrences Spatial functions Adds polygon coverage to a lets.addpoly PresenceAbscence object Adds variables (in raster format) to a lets.addvar PresenceAbscence object Calculates the frequency distribution of a lets.classvar variable within a species’ range Computes correlogram based on Moran's I lets.correl index

lets.distmat Computes a geographic distance matrix

Creates species’ values based on the species lets.field co-occurrence within focal ranges Fits a PresenceAbsence object into a grid in lets.gridirizer shapefile format Creates a matrix summarizing species' lets.maplizer attributes within cells of a PresenceAbsence object Computes species’ geographical range lets.midpoint midpoints Computes pairwise species’ geographic lets.overlap overlap Crops a PresenceAbsence object based on a lets.pamcrop input shapefile

lets.rangesize Computes species’ geographic range size

Filters species’ shapefiles based on its lets.shFilter presence, origin, and season Summarizes variable(s) values within species’ lets.summarizer ranges based on a presence-absence matrix Data download Downloads species’ information from the lets.iucn IUCN RedList online database

123

Downloads species’ habitat information from lets.iucn.ha the IUCN RedList online database Downloads species’ temporal trend in lets.iucn.his conservation status from the IUCN RedList online database

124

Obtain the geographical ranges

of tailless amphibians. (2)

Transform this data into a

presence-absence matrix

(PAM) of 1° of cell resolution

for the whole world. (3) Get the

description year for each species

from the IUCN online database

and summarize it by cells. (4)

Calculate the species

geographical midpoints and

compute the pairwise distance

between them, and use this

information to generate a

Moran’s I correlogram for 200

equidistant classes. (5) Add the

human footprint variable to the

PAM and calculate the

maximum value within each

species’ range, compute the

geographical range size for each

species, and evaluate the effect

of both variables (human

footprint and range size) on Figure 1. Graphical sequence of the analysis to description year variation among species. describe the geographic pattern of description year Functions used at each step of the process are listed in Anura (Amphibia) using the ‘letsR’ package. (1) at the side of each arrow representing these steps.

125

CONCLUSÃO GERAL

Ao fim dessa tese as seguintes conclusões foram obtidas:

Capítulo 1

• O tamanho do corpo em serpentes do Novo Mundo não influencia seu risco de extinção. • Há um viés nas espécies classificadas como DD (Dados Insuficientes), sendo estas menores que as espécies de outras categorias. • Caso as espécies classificadas como DD estiverem na realidade ameaçadas, um padrão inverso ao encontrado em outros vertebrados poderia estar acontecendo, onde espécies de tamanho menor tem um maior risco de extinção. Isso sugere que um viés nas espécies classificadas como DD pode levar a conclusões erradas em análises comparativas do risco de extinção. • O fato de que as espécies menores possuem menos informações é em parte causada pelo fato dessas espécies terem sido formalmente descritas mais recentemente. • Por fim, identificamos uma relação entre o ano em que uma espécie foi descrita e a vulnerabilidade associada. Revelamos um padrão em que as espécies ameaçadas foram descritas em um tempo mais recente do que as não-ameaçadas. Mostramos também que as espécies que não tiveram informação suficiente para uma avaliação (DD) foram descritas em anos mais próximos quando comparado às outras categorias.

Capítulo 2

• Quase todos os grupos de animais avaliados pela IUCN (que tem sua distribuição geográfica disponível) apresentam o mesmo padrão das serpentes, em que quanto mais recente uma espécie foi descrita, mais ameaçada ela está. • A categoria DD apresentou em média as espécies descritas mais recentes, ligando o ano em que uma espécie foi descrita a quantidade de informação associada a ela.

126

• O padrão ‘risco de extinção x ano de descrição’ é causado pelo fato de que espécies descritas mais recentemente possuírem menor tamanho populacional e distribuição geográfica. • As conclusões desse capítulo sugerem que devemos perder as espécies menos conhecidas para a ciência.

Capítulo 3

• Há uma tendência dos gêneros de plantas mais antigos da Península Ibérica Espanhola a serem também mais distintos evolutivamente e apresentar uma baixa taxa de diversificação e riqueza de espécies. • O risco de extinção não é filogeneticamente agregado nas plantas da Península Ibérica Espanhola como um todo, mas sim Andaluzia Oriental. • Não existe uma relação da idade do clado (consequentemente das outras características evolutivas que estavam correlacionadas) com o risco de extinção para toda a Espanha, mas quando avaliamos apenas a Andaluzia Oriental encontramos que espécies mais antigas são mais ameaçadas. • Num olhar mais detalhado, identificamos que tal padrão encontrado na Andaluzia Oriental pode ser explicado pela persistência de espécies ameaçadas e antigas que estavam adaptadas a um clima pré-mediterrâneo. Isso implica que a biogeografia do local determina a relação do risco de extinção com os fatores evolutivos. • Por fim, mostramos que o maior risco de extinção ligado às espécies mais antigas (e distintas evolutivamente), juntamente com a agregação filogenética desse risco na Andaluzia Oriental, deverá causar uma perda filogenética desproporcional na região.

Capítulo 4

• No capítulo 4 disponibilizamos grande parte das ferramentas desenvolvidas para a obtenção, manipulação e análise de dados em um pacote de R, como parte da ideia de que a ciência deve ser aberta e reproduzível. O pacote letsR possui até o

127

dia de hoje 6113 downloads (entre 19 de maio e 2014 e 16 de fevereiro de 2016), o que mostra a amplitude e utilidade dessas ferramentas.

Em resumo, nessa tese mostramos que o tempo considerado da descrição aos dias de hoje pode influenciar diretamente e indiretamente como entendemos o risco associado ao desaparecimento de uma espécie. Também mostramos que o tempo considerado entre o surgimento e os dias atuais, assim como suas variáveis evolutivas correlacionadas, podem ter um papel importante sobre o risco de extinção que observamos atualmente nas espécies, atuando mesmo antes do surgimento da própria espécie. Os dados, análises e procedimentos realizados aqui foram disponibilizados para que futuros estudos ampliem essas conclusões, aprimorem os resultados ou os apliquem a diferentes perguntas.

128