Woody flora of Uruguay: inventory and implication within the Pampean region

Woody flora of Uruguay: inventory and implication within the Pampean region). Adequate knowledge of a flora is fundamental for furthering research in several disciplines. In this context, our work arises from the necessity to update and improve available information on the native woody flora of Uruguay. The goals of this study were to objectively determine the Uruguayan native woody flora (shrub and trees), to evaluate its inventory completeness, and to explore its similarity with neighboring regions. By analyzing the Uruguayan flora we produced working definitions of shrubs and trees, and obtained a list of 313 species (57 families, 125 genera). Using 7418 distribution records, we generated cumulative species richness curves to estimate maximum species richness at national and local scales. We conclude that the completeness at the national level is high (89-95%), but at the local scale is lower and quite heterogeneous. There are still large areas without data or with little information. We found that comparatively, the similarity between Uruguayan tree species and the Province of Entre Ríos (Argentina) was high, while similarity with the Province of Buenos Aires (Argentina) was medium, and with Rio Grande do Sul (Brazil) was low. In conclusion, richness of tree and shrub species of the Uruguayan flora is greater than expected for a grassland region, and the differences found in the similarity index with the neighboring woody floras will stimulate future studies to reevaluate phytogeographic schemes of the region.


Introduction
Floristic studies and inventories attempt to document the biodiversity of a geographic region (Funk 2006).A well--produced flora constitutes a fundamental tool for research in many disciplines, such as ecology, biogeography, evolution and ethnobotany (Mori 1992;Funk 2006;Palmer et al. 1995), and provides a major source of information for conservation, land-use planning and sustainable use of ecosystems (Mori 1992;Palmer et al. 1995;Prance et al. 2000)."Taxonomic impediment" is the critical lack of taxonomic knowledge, and has been considered a major obstacle for conservation and sustainable use of biodiversity (Schnak & López 2003;Crisci 2006;Funk 2006).In this context, our work arises from the necessity to update and improve the knowledge of native woody flora.
Enhanced knowledge of the woody flora of Uruguay will contribute substantially to a more thorough understanding of Uruguayan phytogeography.Uruguay has traditionally been considered a part of the Pampean Province, a region characterized by extensive grasslands or grass-dominated steppes (Cabrera & Willink 1973;Morrone 2001).However, a recent study (Brussa & Grela 2007) suggests that the number of tree and shrub species in the flora might be greater than that expected for the Pampean region.Trees and shrubs are characteristic elements of forests and shrublands, respectively, and represent 12% of all vascular plants in Uruguay, which add up to approximately 2750 species (Alonso & Bassagoda 2002).In Uruguay, the surface occupied by native forests is approximately 820,000 ha (MGAP 2011), which represents almost 4% of the continental territory.The main types of forests present in the country are gallery forests, hill forests and open forests (Chebataroff 1960;Del Puerto 1987;Alonso & Bassagoda 1999).Regarding shrublands, there is scarce information on the area occupied by such woody vegetation.
The main goals of this study were: (1) to determine the native woody flora (shrubs and trees) of Uruguay, (2) to evaluate its inventory completeness, and (3) to explore its floristic similarity with the neighboring regions.

Material and methods
We first reviewed the literature to search for parameters that could provide precise and objective definitions for shrub and tree, applicable for the proposed aims.This step was necessary due to the lack of clear criteria for elaborating a list of shrubs and trees species from the main bibliography of woody plants of Uruguay.
Definitions of trees are common in the literature.Font Quer (1957) defines a tree as a "woody plant, at least 5 m height, with a single stem named trunk (…)".According to Smith et al. (2004) a tree is "an erect, usually single--stemmed, woody plant 5 cm or more dbh (diameter at breast height); some trees may have multiple trunks but at least some of them are 5 cm or more in diameter." Sobral et al. (2006) refer to trees as "erect plants, with well defined and woody stem, with a diameter more than 5 cm measured at 1.30 m above the ground".Regarding shrubs, definitions are also relatively common in the literature.Font Quer (1957) defines a shrub as a "woody plant, less than 5 m height, without a main stem, because it is branched at the base".Smith et al. (2004) define a shrub as "a woody plant that is branched at the base or unbranched but less than 2 m tall".In order to include tree-like plants such as tree ferns or some conspicuous monocotyledons, other authors (Ricker & Hernández 2010) have broadened the definition of trees as: "trees and tree-like plants are defined broadly as perennial, self-supporting plants with an adult height of at least 5 m (without ascending leaves or inflorescences), and with one or several, erect stems with a diameter of at least 10 cm (measured at 1.3 m above the ground level)".
The definitions mentioned above do not completely fit the characteristics of species that comprise the woody flora of Uruguay, mainly because trees of the region tend to be short in height or have slim trunks.Therefore, we hereby propose pragmatic definitions for trees and shrubs, adapting the concepts of the definitions mentioned above, with the characteristics of the woody flora of Uruguay as follows: 1) tree and tree-like plants are terrestrial or hemiepiphyte plants, perennial, erect, with one or few well defined stems, partial or totally woody, of at least 5 cm in diameter measured at 1.3 m above the ground level and an adult height of at least 3 m; 2) shrubs are terrestrial or hemiepiphyte plants, perennial, erect or scandent, branched at the base or close to it, with a minimum height of 1 m above ground level and with aerial woody stems, persistent throughout the year.
These definitions attempt to be sufficiently broad to cover the greatest number of species that grow as trees in broad sense (trees and tree-like) and shrubs.Some species are impossible to classify definitively as trees or shrubs, due to variation in physiognomy or due to the influence of environmental conditions, particularly in this region (Alonso & Bassagoda 2002).As such, these species are identified in this study as belonging to both life forms.
To produce the list of taxa and generate a record database, we used the taxonomic category of species without considering infraspecific categories.Species were named according to Tropicos, botanical information system of the Missouri Botanical Garden -www.tropicos.org (2011) and to the list of Espécies da Flora do Brasil 2011 http:// floradobrasil.jbrj.gov.br/2011.The family classification follows APG III (2009).We used a total of 7418 herbarium specimens from MVFA (4504 specimens), MVJB (2145 specimens), MVM (732 specimens), and MVHC (23 specimens), whose determinations were made by Uruguayan specialists from the consulted herbaria.Specimens with doubtful identifications and/or uncertain localities were excluded.Herbarium acronyms follow Index Herbariorum (Holmgren & Holmgren 1998).Location and other plant trait data were taken from specimen labels.

Data analysis
To analyze the geographic distribution of taxa included in this study, each record was geo-referenced, according to a grid system of 302 quadrants (approximately 22 x 30 km) at the scale 1:50.000(SGM 1990).This information was integrated into a Geographic Information System (GIS).
The inventory completeness for tree and shrub species was assessed at the national level, using all the species included in our list.We first created species accumulation curves for the complete dataset of Uruguay.Then, we fitted the non-parametric model Chao 2 to estimate maximum species richness (with confidence interval of 95%) using EstimateS 8.0 (Colwell 2006).
To assess the adequacy of sampling effort for a complete inventory of the woody flora at a local level (i.e., quadrant--scale) we estimated optimal sampling effort by analyzing the relationship between recorded richness and sampling effort (number of records).A moving average smoothing (lag=10) was fitted to this relationship, and the optimal sampling effort was considered the point where the curve achieved an asymptote.Beyond this point, the observed species richness does not increase consistently with sampling effort, suggesting the complete species assemblage has been registered.The estimation of optimal sampling effort was conducted using floristic regions based on woody plant distributions proposed by Grela (2004).The use of floristic regions takes into account the heterogeneity of flora and vegetation in the country, because areas with different diversity and/or woody vegetation surface may require different sampling effort for proper measurement.Grela (2004) identified two floristic regions with outstanding woody diversity: Western and Eastern Uruguay, the latter with two main subregions (referred here as "Eastern A" and "Eastern B").The rest of the country (mainly central sectors) had lower woody species richness and was considered here the "Central region".The latter comprises 194 quadrants, while the Western covers 40 quadrants, the Eastern A, 22, and the Eastern B, 44.In addition, considering the optimal number of records by region, we calculated the percentage of quadrants sufficiently sampled in relation with the total quadrants of each region.This percentage involves the quadrants with a number of records same as or higher than the optimal number.
To evaluate tree richness similarity between Uruguay and neighboring regions, we used our definition of tree as the criteria to elaborate the species list and the species--region matrix.We analyzed the similarity, using the Jaccard Index, between different regions at species level, in particular for comparisons between Uruguay and the Entre Ríos and Buenos Aires Provinces in Argentina as well as the state of Rio Grande do Sul in Brazil.Information about tree species richness in these regions was obtained and adapted to the proposed criteria from Zuloaga et al. (2008) and Sobral et al. (2006).For all statistical analyses we used the software PAST 2.08 b (Hammer et al. 2001).

Cumulative sampling effort distribution
The accumulated richness curve shows an asymptote at the national scale approaching the maximum number of species (N = 313).The Chao 2 curve also shows an asymptotic pattern, indicating convergence of the estimation procedure (Fig. 1).Estimated maximum richness (mean) based on the Chao 2 model was 328.9 species, almost 16 species more than observed richness (313), indicating the degree of completeness is 95.1%.However, considering the upper bound of the confidence interval, maximum species richness could reach 351.9 species, and thus there could still be 39 unknown species, resulting in a degree of completeness of 88.9%.
The recorded richness per quadrant plotted against number of records shows an asymptotic curve for each region (Fig. 2).In all cases in the range of low records per quadrant, recorded richness increases proportionally with sampling effort; however, beyond a certain value (variable by region) their relationship becomes independent.This suggests that the optimal sampling effort is approximately 86 records per quadrant in the Western region, 56 records per quadrant in the Central region, and 183 and 96 records 3 Randia sp.This species has been traditionally considered as Randia armata (Sw.)DC., but currently is considered a doubtful classification (Marchesi pers.com.).
Continues.per quadrant in the Eastern A and the Eastern B regions, respectively (Fig. 2).The geographic distribution of records within each region is quite heterogeneous (Fig. 3), with some areas showing high concentration of records and others with a few to no records at all.The proportion of sufficiently sampled quadrants ascend to 7.5% in the Western region, 7.7% in the Central region, 13.6% in the Eastern A region and 9.1% in the Eastern B region.Quadrants without records (i.e., information gaps) represent 10%, 29.4%, 9.1% and 18.2% of each of these regions, respectively, and 7% of the whole country.

Similarity between tree species of Uruguay and neighboring regions
Uruguayan tree species richness was similar to that registered for the Entre Ríos Province, which was higher than that of the Buenos Aires Province but much lower than that of Rio Grande do Sul State (Table 2).Indeed, Rio Grande do Sul has the highest values for families, genera and species richness of the four compared regions (Table 2).Tree richness per surface unit indicates values for Uruguay that are four times higher than that of Buenos Aires, and about a half of that of Rio Grande do Sul (Table 2).The similarity analysis among regions shows greater similitude of Uruguayan tree flora with Entre Ríos than any of the other regions and the lowest similarity with Rio Grande do Sul (Table 3).

Discussion
This work generated precise definitions of trees and shrubs appropriate for the Uruguayan context that, despite being arbitrary, attempt to avoid the ambiguity and possible confusion when compiling species lists.The difference between the species list reported here and the list generated by Lombardo (1964) is easily explained by new records of species for the Uruguayan flora reported more or less continuously during the last 46 years (Table 4).With regards to Brussa & Grela (2007), the number of species is similar to that presented here, but there is a considerable difference (24%) in species composition.This difference is mainly due to the lack of explicit criteria for inclusion of species in the aforementioned publication, resulting in the exclusion of   species when our definitions are applied (Table 4).In addition, the work of Brussa & Grela (2007) emphasizes species from Tacuarembó and Rivera, thus the current list includes several species elsewhere in the country not present in that work.We also excluded some species present in Brussa & Grela (2007) because their presence in Uruguay has not been confirmed or for lack of taxonomic updates or knowing the exact location of specimens.
Alternatively, the information for some taxonomic groups (e.g., Baccharis L., Croton L., Opuntia Mill.) still remains incomplete due to the high taxonomic complexity and elevated diversity of Baccharis and Croton in Uruguay, as well as the lack of specimens of Opuntia.So, new species may be added to the list of trees and shrubs present in the country.Despite the fact that the number of taxonomic works focused on woody species has increased in the last few years, there are rather few, thus stressing the need for further studies that address untreated families and genera.
From the standpoint of inventory completeness, our results suggest that the accumulated knowledge of trees and shrubs of Uruguay is fairly complete, with only 11% of the potential number of woody species yet to be registered.The floristic regional completeness within Uruguay allows a more accurate approximation of the sampling effort needed in areas of the country with different environmental characteristics.In all regions there is more or less heterogeneity of the sampling effort, with areas with scarce or no information.The Eastern A region has the highest degree of completeness, while the Central region has the highest proportion of gaps.This situation results from a lack of planned and systematic sampling, a common problem in many areas (Spichiger et al. 2004).Instead, ease of access and landscape attractiveness has generated the high concentration of sampling in some areas while neglecting others.The low degree of completeness in the Central region is echoed in other biological groups (Brazeiro et al. 2008;Canavero et al. 2010).Apart from the Central region the analysis presented here allowed us to identify other areas with scarce completeness, whose study should be marked as a priority in future studies and surveys.The comparison of Uruguayan tree species richness with neighboring regions offers further support for the idea that there are differences between Uruguay and the Province of Buenos Aires flora (Chebataroff 1942;1960;Cabrera & Willink 1973;Del Puerto 1987;Alonso & Bassagoda 2002;Grela 2004), which is consistent with the demarcation of a Uruguayan District inside the Pampean Province as proposed by Cabrera & Willink (1973).
The transitional nature of the Uruguayan vegetation and flora explains the relatively high tree and shrub species richness, especially when the small continental land area of the country and particularly the scarce native forest surface are taken into account.The Uruguayan territory represents the southernmost and southeastern boundaries for the distribution of many species, with some showing restricted distribution within the country (Alonso & Bassagoda 2002;Brussa & Grela 2007).
In conclusion, despite the fact that Uruguay lies within a region dominated by grasslands, the country has a surprisingly high richness of tree and shrub species and its woody flora shows considerable differences with those of neighboring regions.The difference with the typical Pampean flora analyzed is remarkable, represented by Buenos Aires.Detailed future studies of the woody flora and vegetation will allow re-evaluating the proposed phytogeographic schemes and shedding further light on the plant distribution patterns for the region.

Figure 2 .
Figure 2. Relation between observed richness per quadrant and the sampling effort (number of records) for each region.The solid curve was smoothing by moving average (lag: 10).The dotted line shows the value where the sampling effort starts the saturation.

Figure 1 .
Figure 1.Cumulative richness of tree and shrub species in Uruguay based on the distribution of records.The estimated richness was calculated with Chao 2 index.IC = confidence interval of 95%.

Figure 3 .
Figure 3. Geographic distribution of sampling effort by quadrant in Uruguay for each region.

Table 4 .
Number of tree and shrub species and families, proposed by different authors and by the present work.