A list of land plants of Parque Nacional do Caparaó, Brazil, highlights the presence of sampling gaps within this protected area

Abstract Background Brazilian protected areas are essential for plant conservation in the Atlantic Forest domain, one of the 36 global biodiversity hotspots. A major challenge for improving conservation actions is to know the plant richness, protected by these areas. Online databases offer an accessible way to build plant species lists and to provide relevant information about biodiversity. A list of land plants of “Parque Nacional do Caparaó” (PNC) was previously built using online databases and published on the website "Catálogo de Plantas das Unidades de Conservação do Brasil." Here, we provide and discuss additional information about plant species richness, endemism and conservation in the PNC that could not be included in the List. We documented 1,791 species of land plants as occurring in PNC, of which 63 are cited as threatened (CR, EN or VU) by the Brazilian National Red List, seven as data deficient (DD) and five as priorities for conservation. Fifity-one species were possible new ocurrences for ES and MG states. New information "Parque Nacional do Caparaó" houses 8% of the land plant species endemic to the Brazilian Atlantic Forest, including 6% of its angiosperms, 31% of its lycophytes and ferns and 14% of its avascular plants. Twelve percent of the threatened species listed for the State of Espírito Santo and 7% listed for the State of Minas Gerais are also protected by PNC. Surprisingly, 79% of the collections analysed here were carried out in Minas Gerais, which represents just 21% of the total extension of the Park. The compiled data uncover a huge botanical collection gap in this federally-protected area.


Abstract Background
Brazilian protected areas are essential for plant conservation in the Atlantic Forest domain, one of the 36 global biodiversity hotspots. A major challenge for improving conservation actions is to know the plant richness, protected by these areas. Online databases offer an accessible way to build plant species lists and to provide relevant information about biodiversity. A list of land plants of "Parque Nacional do Caparaó" (PNC) was previously built using online databases and published on the website "Catálogo de Plantas das Unidades de Conservação do Brasil." Here, we provide and discuss additional information about plant species richness, endemism and conservation in the PNC that could not be included in the List. We documented 1,791 species of land plants as occurring in PNC, of which 63 are cited as threatened (CR, EN or VU) by the Brazilian National Red List, seven as data deficient (DD) and five as priorities for conservation. Fifity-one species were possible new ocurrences for ES and MG states.

Introduction
The Atlantic Forest is recognised worldwide for its high biological diversity, high rates of endemism and great threat (with more than 70% of its original area devastated) and which is considered one of 36 global biodiversity hotspots (Marques and Grelle 2020). It has been estimated that this phytogeographic domain contains 17,776 species of land plants with an endemism rate of 56% (Flora do Brasil under construction 2020). Only 28% of the Atlantic Forest original area remains and its area consists mainly of edge-affected or secondary vegetation disconnected from larger remnants due to the long and intense disturbances it has experienced throughout its history (Rezende et al. 2018). There are 1,437 protected areas (PAs) in the Atlantic Forest, including strictly-protected areas (477) and areas managed for sustainable use (960) (MMA 2020). However, only 30% of the total vegetation cover is located within PAs, of which 9% are strictly protected (Rezende et al. 2018) and little is known about the plant species that are protected within them (Oliveira et al. 2017).
In the past 20 years, virtual tools have become more popular by sharing information from herbaria all over the world to contribute and to facilitate voucher identification, such as the Global Biodiversity Information Facility (GBIF; http://www.gbif.org), REFLORA (http:// reflora.jbrj.gov.br) and INCT-Splink (http://inct.splink.org.br). The availability of biodiversity data on the internet has not only increased communication amongst herbaria, it has also provided information for documenting biodiversity and its distribution in space and time and, thus, has served as a backbone for developing environmental policies (Maldonado et al. 2015, Nualart et al. 2017, Oliveira et al. 2017). In addition, as the Atlantic Forest is one of the best-sampled phytogeographic domains in Brazil, the use of online databases can assist in documenting biodiversity and revealing temporal, spatial and taxonomic gaps in knowledge (Oliveira et al. 2017, Colli-Silva et al. 2020. "Catálogo de Plantas das Unidades de Conservação do Brasil" (https://catalogo-ucsbrasil.jbrj.gov.br/) is a digital platform created in 2018 with the aim to host lists of plant species that occur within Brazilian protected areas. Currently, the catalogue contains plant lists for five PAs from different Brazilian phytogeographic domains, such as the Amazon, Caatinga and Atlantic Forest. The plant list for "Parque Nacional do Itatiaia" (PNI) was the first made available on this platform , while the list for "Parque Nacional do Caparaó" (PNC) was the most recent . In addition to access to lists of species and vouchers, the website also provides the threat category of species, information on the occurrence of non-native and native species and a list of priority species for conservation for each protected area. "Parque Nacional do Caparaó" (PNC), created in 1961 and effectively implemented in 1979, aims to protect areas of the Atlantic Forest domain (Santos 2004, ICMBio 2015. The first scientific expeditions to the area now occupied by the PNC were carried out by Wilhelm Schwacke in 1880 and by the Belgian-Brazilian Mission in 1922 (Massart et al. 1929). The first floristic inventory, however, was carried out by Alexander Curt Brade in 1942 and this recorded 259 species of plants of 59 families (Brade 1942). Other Botanists who contributed significantly to the knowledge of the local flora were Alexandre Salino, Braz AP Cosenza, Leopoldo Krieger, Lúcio de Souza Leoni and Vinicius Castro Souza (Santos 2004). Despite this intense collection effort, knowledge of the flora of PNC is dispersed in numerous publications on specific groups or families of plants (Aguiar and Marques 2001, Souza and Souza 2002, Romão and Souza 2003, Leoni and Chautems 2004, Mazine and Souza 2008, Leoni 2009a, Leoni 2009b, Forster and Souza 2013, Góes-Neto et al. 2015, Machado et al. 2016, Góes-Neto and Salino 2018, Cardoso et al. 2019, Camelo et al. 2020). There remains, however, a lack of a single source providing access to reliable information about plant species protected by this PA. Here, we provide and discuss additional information about species richness, endemism and conservation in PNC that could not be provided in the plant list for PNC previously published in the catalogue (after ).
To obtain a list of species with currently-accepted nomenclature, we manually selected all specimens identified to the species level, leading to the following: JABOT determined = 3,113, undetermined = 1,074; REFLORA determined = 6,637, undetermined = 2,768; and Splink determined = 7,922, undetermined = 3,141 (Fig. 1). We also removed specimens that had a locality that did not belong to the area covered by the PNC, as well as duplicates (based on the catalogue code, collector name and number and the year in which the sample was collected) (Fig. 1). We then corrected and updated species names and determined their threat categories using the function get.taxa from the flora package ) of R software v. 3.5.3 (R Development Core Team 2019). We used this function to compare the names in our list with those in Flora do Brasil under construction (2020) and determined the threat category for each species according to the Red List Authority for plants in Brazil -CNCFLora (http://www.cncflora.jbrj.gov.br/portal) (Fig. 1). Introduced species were not recovered by the function get.taxa and so these specimens had to be reviewed manually. After these corrections, taxonomists checked the preliminary list of 2,372 species virtually, using images available in the online databases (Fig. 1). When a taxonomist modified a plant species name, at least one specimen of that species was updated in the Herbarium of "Jardim Botânico do Rio de Janeiro" (RB, acronym from Thiers 2020) and its database JABOT, as well as in the REFLORA database, but not in Splink. Infraspecific taxonomic categories were not considered, nor were hybrids.
To assess the most collected sites in the Park, we built a word cloud, based on the name of locations where specimens were collected. We did not include broad locations in the word cloud, such as "Serra do Caparaó" and "Parque Nacional do Caparaó", since they are not informative. We constructed the word cloud using the function wordcloud of the wordcloud package version 2.6 (Fellows 2018). Stages of data cleaning performed in R software to obtain a list of land plants of "Parque Nacional do Caparaó," Brazil, from the online databases.
To evaluate whether species were native or non-native and endemic or non-endemic to Brazil and to assign a threat category, we used information from Flora do Brasil 2020 (under construction; http://floradobrasil.jbrj.gov.br) and the Red List Authority for plants in Brazil -CNCFLora (http://www.cncflora.jbrj.gov.br/portal). In the case of species not included in the Flora do Brasil 2020 database, threat category and origin (native/nonnative) were obtained from taxonomists. We considered as non-native all species indicated as cultivated or naturalised by the Flora do Brasil 2020 database. We also assessed whether species of PNC were on the threatened species lists in the States of Espírito Santo (Fraga et al. 2019) and Minas Gerais (COPAM-MG 2008). We classified species as endemic to the Atlantic Forest when their distribution is restricted to the Atlantic Forest phytogeographic domain. We obtained this information from the Flora do Brasil 2020 database through the function get_domains from the flora package . This classification also included species non-endemic to Brazil as the Atlantic Forest occurs in the countries of Argentina and Paraguay. We classified species as rare in Brazil, based on the publication of Giulietti et al. (2009). We classified species as a priority for conservation when they had a single record collected before 1970 (Briggs and Leigh 1988) and was simultaneously categorised as Critically Endangered (CR), Vulnerable (VU), Endangered (EN) or Data Deficient (DD), according to Red List Authority for plants in Brazil -CNCFLora (http://www.cncflora.jbrj.gov.br/portal). Mean annual rainfall in PNC is 1,481.1 mm, with the greatest rainfall in January (mean 316.7 mm) and the lowest in July (mean 16.2 mm; data are mean rainfall for 1974-2003ICMBio 2015). The mean annual temperature is 11°C, with a minimum of 2.5°C in July and a maximum of 31°C in December (data are mean temperatures between December 2004 and December 2005 at 2,400 m a.s.l.; Rodrigues et al. 2009). The vegetation of PNC comprises different phytophysiognomies, including dense ombrophilous montane forest, dense ombrophilous high-montane forest, seasonal semi-deciduous montane forest and high-altitude grassland (campos de altitude) (Fig. 2b, c, d;ICMBio 2015).

Figure 6.
Native and non-native species to Brazil of major groups of plants that occur in "Parque Nacional do Caparaó," Brazil. Values above the bars indicate the number of species. We found a similar number of non-endemic (900) and endemic (891) species to Brazil (Fig.  7). Most of both the endemic and non-endemic species to Brazil are angiosperms (83% and 61%, respectively), followed by ferns (11% and 19%), avascular plants (4% and 18%) and lycophytes (2% and 2%). Orchidaceae and Asteraceae have the highest number of endemic species to Brazil (92 and 74, respectively), followed by Melastomataceae (62), Bromeliaceae (41) and Myrtaceae (37) (Fig. 8). The PNC protects 785 endemic species in the Atlantic Forest domain (Fig. 7), most of which are angiosperms (64.3%), followed by ferns (20.3%), avascular plants (12%) and lycophytes (3.4%). Orchidaceae has the highest number of endemic species to Atlantic Forest (67), followed by Dryopteridaceae and Asteraceae (both with 37), Melastomataceae (35), Solanaceae and Polypodiaceae (both with 33) and Bromeliaceae (32).  Orchidaceae having the highest number of threatened species (7) followed by Bromeliaceae (6), Asteraceae (5), Polypodiaceae (4) and Gesneriaceae (3)   Considering the entire flora of PNC, 24 species belonging to 18 families are rare in Brazil. Asteraceae and Orobanchaceae have the highest number of rare species (three) followed by Bromeliaceae and Scrophulariaceae (two each, Table 2). Fourteen families have a single rare species (Table 2). We retrieved five species, four angiosperms and one fern, which are priorities for conservation in PNC (Table 1).
Rare species of plants occurring in "Parque Nacional do Caparaó," Brazil and their respective families.

Families
The full scientific name of the family in which the taxon is classified

Genera
The full scientific name of the genus in which the taxon is classified.

Species
The full scientific name Controlled vocabulary ("yes", "no") The species present only old records?

Conclusions and prospects
The analysis of specimens collected in the PNC allowed us to detect a spatial collection gap. Most (79%) of the specimens analysed were collected in a small portion of the Park (21% of the area of the Park) located in MG. These collections were mainly made in easily accessible places of PNC, such as Vale Verde, Cachoeira Bonita, Vale Encantado, Tronqueira and Terreirão (Fig. 11). Collection bias towards access routes is a common sampling problem that has been reported for various taxonomic groups (i.e. vertebrates, invertebrates and angiosperms) throughout all Brazilian phytogeographic domains and can affect the detection of spatial patterns of species diversity (Oliveira et al. 2016).
Furthermore, amongst the species found in PNC, 445 are described by Flora do Brasil under construction (2020) as occurring in MG, while only 81 species are described as occurring in ES, whereas a total of 1,199 occur in both States (ES and MG; Fig. 12 and Suppl. material 1). The Park also houses only 14% (873 species) of the species, 29% (398) of the genera and 61% (110) of the families of angiosperms listed for ES (Suppl. material 1; Dutra et al. 2015). We did not find information about the state of occurrence for 13 species and 53 species did not appear in Flora do Brasil under construction (2020) as occurring in either ES or MG ( Fig. 12 and Suppl. material 1). Despite the Atlantic Forest being one of the most sampled phytogeographic domains in Brazil ), our results demonstrate that even areas that have been extensively studied can have collection gaps. Thus, deeper knowledge of the flora of PNC requires increased collection effort in the less accessible areas of the Park located mainly in ES. Word cloud of locations where plants of "Parque Nacional do Caparaó," Brazil, were collected. The different colours and sizes of the letters represent the frequency of collections at the localities within the Park; larger font size represents a greater number of collections.
A particularly interesting result was the 53 species in the list of PNC that were not recorded as occurring in ES or MG. It may be that these species are new occurrences for MG or ES. One recent study, involving the family Araceae in PNC, reported Xanthosoma maximilianii Schott and Philodendron acutatum Schott of these 53 species, as new occurrences for ES and MG, respectively (Camelo et al. 2020), emphasising the importance of more in-depth assessments of the occurrence of the other 51 species.
The species list for PNC, presented here, was prepared using information acquired from online databases and validated by taxonomists. This method is considered good practice for estimating species diversity in the "era of big data" (Maldonado et al. 2015). The retrieval of information contained in the online databases revealed that PNC houses 10% (1,699 species) of the species recognised for the Brazilian Atlantic Forest (i.e. 17,776 species; Flora do Brasil under construction 2020) and 8% (785) of the species endemic to this domain (i.e. 10,024 species; Flora do Brasil under construction 2020), with 6% (505) of angiosperms, 31% (186) of lycophytes and ferns and 14% (94) of avascular plants. In conclusion, the species list of PNC allowed us not only to identify the species that occur in this protected area, but also to identify gaps in knowledge that can help direct the allocation of future collecting efforts.