Richness of Cerrado Woody Species Engaged in Ecological Restoration in the Brazilian Federal District

This study has evaluated the richness of Cerrado woody species engaged in ecological restoration in the Brazilian Federal District (BFD). A survey gathered information on plant species traded by local nurseries, species recommended in restoration plans (PRADs), species effectively introduced in areas under restoration, and species present in preserved fragments of Cerrado. Results summed 566 Cerrado woody species from 80 botanical families of which 171 species were traded by local nurseries, 277 were recommended in PRADs, 190 were effectively used in restoration projects, and 434 species were sampled in fragments of native Cerrado. We found low similarity between species composition available in nurseries, recommended in PRADs, used in restoration projects and present in preserved fragments of native Cerrado. Such results indicate a poor connection between steps related to the selection of native woody species that make up initial plant communities on sites under ecological restoration.


INTRODUCTION
The Cerrado biome houses more than 11,000 species of vascular plants and is the richest savanna in plant species in the world (Mendonça et al., 2008). This biome has undergone severe degradation from the 1960s (Rada, 2013) mainly to support agriculture, urbanization, and mining activities (Beuchle et al., 2015;Klink & Machado, 2005;Sano et al., 2010;Spera et al., 2016). As a result, only 54% of the original area remains under natural vegetation cover (Brasil, 2015).
Huge environmental liabilities in Brazilian biomes have triggered the demand for ecological restoration plans and projects, which aim at implementing strategies to rehabilitate natural ecosystems, environmental services, and ecological sustainability on degraded sites (Chazdon, 2008). As such, restoration projects intend to recover ecological and structural characteristics of ecosystems close to the previous original conditions (Palmer et al., 2016). Many techniques have been used to achieve restoration goals, such as natural regeneration, assisted regeneration, direct seeding, nucleation, and tree plantation, which is the most common and traditional practice of ecological restoration (Reis et al., 2010).
When projects rely on tree plantation, species richness and composition are vital characteristics of plant communities that will start up ecological succession on sites under restoration (Crouzeilles et al., 2017;Rodrigues et al., 2009;Siqueira et al., 2015). However, many projects that opted for tree plantation have selected a limited number of woody species to compose initial plant communities (Barbosa et al., 2003;Brancalion et al., 2013;Corrêa et al., 2015;Durigan et al., 2010;Rodrigues et al., 2009). Failures in restoration projects have often been attributed to low species richness and low diversity of initial plant communities (Barbosa et al., 2003).
Based on such a scenario, our objectives were to evaluate the richness of Cerrado woody species that has been recommended, available, and used for ecological restoration in the Brazilian Federal District and analyze its implications.

Study area
This work was developed in the Brazilian Federal District (BFD), which is located on the Brazilian Central Plateau (Oliveira & Pompermayer, 2012). BFD's altitude ranges from 1,000 m to 1,200 m (Martins et al., 2004), local climate is Tropical Savanna (Aw, in Köppen-Geiger classification) with dry winters, rainy summers, and an annual rainfall mean of 1,500 mm (INMET, 2018). All the fourteen Cerrado phytophysiognomies occur in the BFD (Walter, 2001) and 38.0% of BFD's territory was originally covered by savanna formations, 43.2% by grassland formations, and 18.8% by forest formations, from which 5% were gallery forests (UNESCO, 2002).

Data gathering
Data were gathered from (1) plant nurseries, (2) restoration plans (PRADs), (3) academic works on plant species introduced in areas under restoration, and (4) species naturally present in preserved fragments of Cerrado. These four categories of sources were surveyed until sampling sufficiency was achieved for each category. Names of Cerrado woody species from 21 nurseries, 35 PRADs, 21 implemented PRADs, and 10 fragments of Cerrado were organized for this work (Figure 1). Lists of woody species recommended in PRADs were compiled at the local environmental agency library (Instituto Brasília Ambiental -IBRAM) and at the website Biblioteca Digital (IBRAM, 2018). Taxa names were updated online as per the nomenclature of the Missouri Botanical Garden (MOBOT, 2016). Botanical families were organized according to The Angiosperm Phylogeny Group et al. (2016).

Data analysis
Rarefaction curves (Colwell et al., 2012) periodically tested sampling sufficiency for each of the four surveyed categories by using the software R Core Team version 3.5.1 (2017). Plant species within surveyed categories were compared by using Vegan package, according to Ugland et al. (2003), Colwell et al. (2004), and Kindt et al. (2006). Interpolated and extrapolated estimates of species richness were run at 95% confidence level as permutation allows drawing average curves of species accumulation and their empirical confidence intervals (Schilling et al., 2012). Bootsrap species richness estimator was used for species analysis and categorical data matrices were generated from the occurrence of the presence of species in each surveyed category.
Species origin was checked according to Mendonça et al. (2008) and Cerrado native species were categorized according to their natural occurrence in Cerrado phytophysiognomies (Ribeiro & Walter, 2008). Data were organized on a table for summarizing the total number and the percentage number of Cerrado woody species engaged in each of the four surveyed categories.

RESULTS AND DISCUSSION
The survey of Cerrado woody species in 21 nurseries, 35 PRADs, 21 implemented PRADs, and ten preserved fragments of Cerrado in the Brazilian Federal District (BFD) were enough for reaching stability tendency on rarefaction curves ( Figure 2). Procedure on sampling sufficiency adopted in this study emphasizes the asymptotic response to successive samplings ( Figure 2) since plant species data from tropical ecosystems do not usually achieve inflection points on rarefaction curves (Corrêa et al., 2015;Schilling et al., 2012) as also shown in our work ( Figure 2). This survey summed 566 Cerrado species from 80 botanical families, which account for 13.2% of the Cerrado's vascular plant species identified so far (Mendonça et al., 2008). There were 171 species from 45 families traded by local nurseries, 277 species from 64 families recommended in PRADs, 190 species from 52 families effectively used in restoration projects, and 434 species from 72 families were sampled in fragments from native Cerrado (Table 1)    Kielmeyera lathrophyton Saddi sec.

-23
Botanical   The number of woody species traded by local nurseries made up 39% of the same found in native fragments of Cerrado and it may be insufficient to meet the demand of restoration plans for achieving rich and diverse plant communities. However, this scenario is better than the one found by Oliveira et al. (2017), who evaluated the availability of native species saplings in nurseries settled in the Rio Grande catchment area, Minas Gerais (Brazil) and found a species richness lower than 10% compared to the regional native vegetation. Cerrado is the species-richest savanna in the world (Mendonça et al., 2008) and the relatively low species richness available in nurseries (39%) is attributed to difficulties in collecting seeds from a wide range of native species and the poor knowledge on germination and growth of many native plant species (Oliveira et al., 2016;Oliveira et al., 2017;Santos & Queiroz, 2011). The main obstacles for plant sapling production of Brazilian native species are seed shortage (80%), difficulties in trading plant saplings (75%), and poor training for the management of native species (65%) (Silva et al., 2017).
Species richness recommended in the surveyed PRADs achieved 63.8% of that naturally present in fragments of Cerrado. However, restoration plans represent only the intention of setting up highly-diverse plant communities that will trigger ecological succession in degraded areas (Corrêa et al., 2015). Examined PRADs showed superficial and incomplete approaches to the problems intended to tackle, as some plans mostly swerved around real characteristics of sites to be restored and many proposed plant species were not adequate to them. Therefore, some PRADs were rather instruments to comply with environmental laws than to outline effective ecological restoration (Lima et al., 2006). There were lists of activities and plant species in these plans that did not match the availability of sapling species traded in BFD nurseries. Sánchez (2010) pointed out three major problems associated with PRADs: i) they usually are improperly drawn up and it results in unsatisfactory restoration when applied in practice; ii) they should be periodically updated; iii) proposed measures in PRADs are vague, generic, and difficult to check.
Low average of species richness recommended in PRADs and in executed PRADs may be a result of low availability of native species in individual nurseries, although the pool of 21 surveyed nurseries in BFD traded 171 Cerrado woody species as a whole. Thus the range of 20-24 species introduced as initial plant communities on restoration sites is not reasonable because plant species for a given PRAD can be purchased from more than one nursery. Surprisingly, we found 190 Cerrado wood species on sites where PRADs had been executed and such figure suggests that some species could have come from elsewhere besides local nurseries. The introduction of species from other populations may lead to genetic contamination, extinction of local populations, and loss of genetic biodiversity, which opposes one of the ecological restoration goals. Yet, introduction of tree saplings from distinct ecological regions brings back genes that natural selection had already banned from the receiving area or genes previously inexistent in it (Durigan et al., 2010).
Species-rich plant communities may guarantee restoration success as some studies point out that increases in ecosystem functions follow increases of species richness (Cardinale et al., 2007;Solan et al., 2009). Barbosa et al. (2003) found 355 native species in 30 plant nurseries in São Paulo State, Brazil, and an average of 30 native woody species in executed PRADs. The authors have attributed the low species richness on sites under restoration to the low availability of species in local nurseries. By comparison with our data, it seems that a low number of plant species available in individual nurseries have translated into low species richness in areas under restoration (Barbosa et al., 2003).
Qualea grandiflora Mart. was the most frequent species found in preserved fragments of Cerrado in BFD and it was present in 80% of the surveyed sites. Tabebuia roseoalba (Ridl.) Sandwith was the most frequent species available in local nurseries and it was sold by 86% of the surveyed traders. Caryocar brasiliense Cambess. was the most recommended species in PRADs and appeared listed in 49% of them. Finally, Copaifera langsdorffii Desf. was the most frequent species effectively introduced in degraded areas and it was sampled in 62% of sites under restoration. Such a figure reflects the poor connection between the stages necessary for achieving a sound ecological restoration: reference ecosystem (Cerrado fragments), planning (PRADs), necessary support (nurseries), and execution of restoration projects.
The low number of pioneer species in areas of executed PRADs may also be a problem as only 23.7% of plant species in such areas are pioneer species (Table 1). São Paulo State regulation SMA 32/2014 requires 40% of pioneer species to compose initial plant communities on sites that will undergo ecological restoration. Pioneer species usually grow faster than plant species of advanced ecological stages (Durigan et al., 2010) and it hastens the development of vegetation cover, which is an essential step towards the restoration of ecosystems (Corrêa et al., 2018). Another critical issue on BFD ecological restoration refers to the widespread use of forest species in areas of previously inhabited savanna formations. Such practice will likely lead succession towards the formation of forest ecosystems (Overbeck et al., 2013;Parr et al., 2014;Veldman et al., 2015).
Production of woody saplings from many different native species is a factor that currently limits ecological restoration in many parts of Brazil (Silva et al., 2017). There is currently a lack of knowledge on the production of plant saplings for several Cerrado native species (Barbosa et al., 2003;Oliveira et al., 2016;Oliveira et al., 2017;Santos & Queiroz, 2011). As a result, it is rather difficult to find a broad sort of woody species in commercial nurseries (Oliveira et al., 2016). Seed collection and appropriate germination protocols for Cerrado species are other limitations for ecological restoration (Viani & Rodrigues, 2007), although there are already studies on these issues (Young et al., 2005).
Besides the difficulties to produce plant saplings from Cerrado species and the low species richness in BFD nurseries and PRADs, our study shows the detachment between species composition along the line nurseries, PRADs, and executed-PRADs, as only 22.9% of species were common to these three categories.

CONCLUSION
Cerrado woody species available in nurseries established in the Brazilian Federal District (BFD) made up 39% of the species richness found in native fragments of Cerrado as a whole. However, species richness found on sites under restoration falls to 5.5% of it on average. Total number of plant species traded in nurseries (171) can support plant communities richer in species than the ones recommended in PRADs (20 on average) and found in areas under restoration (24 on average). Restoration plans should therefore rely on various nurseries to increase species richness in initial plant communities.
There was a higher number of Cerrado species recommended in PRADs (277) than available in BFD nurseries (171) or growing in areas of executed PRADs (190). Such a figure portraits the unrealistic nature of the surveyed restoration plans.