Palm Forest Landscape in Castillos ( Rocha , Uruguay ) : Contributions to the Design of a Conservation Area

Butia palm forests are considered unique due to their aesthetic value, high biodiversity level and historical, archaeological and cultural value. The lack of regeneration of butia palms caused by cattle overgrazing and natural grasslands replacement by agriculture endangers these palm forests. The aim of this work is to provide information for the proposal of a conservation area in this rural landscape within the framework of sustainable development. This work was developed within the context of a Geographic Information System with thematic information on palm forest density levels, soils, land use aptitude and rural registers. Field surveys were conducted to record the presence of plant species and genetic resources in different vegetation units. The main category of soil cover was natural grasslands, followed by native forests, wetlands and palm forests. Palm forests grow mainly on soil units with poor to rather poor drainage and on arable or non-arable lands under special conditions. We identified 212 farms where the palm forest is found, a high diversity of vegetation units, the presence of 302 native taxa of plants and a significant number of plant genetic resources for various uses. For the designation of a conservation area, we propose 20 priority farms that will ensure that palm forests are well represented regarding ecosystem diversity. We consider the creation of a Protected Landscape, a State Park or a GIAHS to be the best alternatives so as to protect this unique multifunctional landscape.


Introduction
Butia palms (Butia odorata (Barb.Rodr.)Noblick) are without a doubt a plant genetic resource with multiple uses that have shaped a biocultural and multifunctional landscape since pre-historical times (López Mazz & Pintos, 2001;Dabezies, 2011;Rivas & Barbieri, 2015), in the sense proposed by Agnoletti & Rotherham (2015).Butia odorata (butia) palm forests form a landscape characterized by the presence of butia palms with density levels of between 50 and 600 palms per hectare over natural herbaceous grasslands.These forests are found exclusively in the Pampa biome, in the south of Brazil and Uruguay.In the Department of Rocha (Uruguay is subdivided in 19 administrative regions known as departments), in the southeast of Uruguay, two large areas of palm forests are found: the palm forests of Castillos and San Luis, which cover an area of almost 70,000 hectares (Rivas, 2005).These palm forests are considered to be unique because of their landscape and biodiversity value, and also due to the associated historical, archaeological and cultural values (PROBIDES, 1995;López Mazz & Pintos, 2001;Geymonat & Rocha, 2009;Rivas & Barbieri, 2015).The Castillos palm forest is located in the middle plains of the Laguna Negra (Black Lagoon) basin, an area characterized by high environmental heterogeneity as a result of a topographic gradient from the hills to the lowlands (PROBIDES, 1999).This territory is part of the Biosphere Reserve of Bañados del Este (Eastern Wetlands) (MAB-UNESCO), with the presence of forests and bushes, natural fields or grasslands with varied physiognomies and a rich botanical composition as well as wetlands and coastal vegetation (PROBIDES, 1999;Pezzani, 2007).
While in Uruguay, the destruction of natural palm forests is forbidden (Law No. 9872 of 1939, as amended by Law No. 15,939 of 1987), this does not ensure their conservation.Most palms of the forest are more than two or three hundred years old, and there are almost no young plants to ensure the forest's survival.
The lack of regeneration poses a serious threat to their preservation.These processes take place in lands which are owned exclusively by private owners for agricultural use.In the Castillos palm forest, the main activity is cattle farming on natural grasslands and, more recently, extensive agriculture.Overgrazing is a common practice in which cattle often eats and crushes young butia seedlings, causing their death.On the other hand, agriculture prevents seeds from germinating and seedlings from growing (Rivas, 2005;Rivas, 2013).
The Guidelines for Land Use Planning and Sustainable Development of the Government of Rocha state that "the aim of the development policy is to achieve a virtuous link between conservation and sustainable use of environmental values".In particular, "the conservation of palm forests in a production environment that makes its reproduction viable" is quoted as an expected result, while "promoting conservation actions for palm forests within productive land uses" appears as one of the plan's main actions.
In this sense, the concept of in situ conservation (United Nations, 1992, 2002;FAO, 1996FAO, , 2012) ) applies to the butia palm forest landscape.The framework for the establishment of a conservation area and the design of a management plan is that of sustainable use of biodiversity and their associated traditional knowledge (Hawkes et al., 1997;Maxted et al., 1997;Perrino et al., 2006;Iriondo et al., 2008;Rivas et al., 2010;Rajpurohit & Jhang 2015;Rivas & Condón, 2016).On the other hand, the multifunctional approach on landscapes (Cullotta & Maetzke, 2009;Taylor et al., 2010;Reyers et al., 2012;Dale et al., 2013) takes into consideration the advantages of landscape heterogeneity and allows us to meet specific ecological, productive and cultural objectives in order to improve the condition of the land and the quality of life of its inhabitants (Taylor et al., 2010, Brussaard et al., 2010).
The overall objective of this study is to contribute with recommendations for the establishment of a conservation area and the formulation of a management plan that helps to preserve butia forests within the context of a multifunctional rural landscape.The specific objectives are to identify and quantify the various categories of land cover, to survey plant communities, species and plant genetic resources, the characterization of the environmental conditions in which palm forests are found and their distribution within commercial farms.GIS (Geographic Information Systems) and ecogeographic surveys (Guarino, 1995;Magos Brehm et al., 2008;Parra-Quijano et al., 2012;Berlingeri & Crespo, 2012) are seen as appropriate tools so as to develop these objectives.

Area under Study
The studied territory was defined based on the distribution of the Castillos palm forest (Rocha, Uruguay) and its area of influence (Figure 1).The coordinates of the area are: 33˚53'14.29"S,53˚59'14.69"W-northwest;33˚52'38.67"S,53˚36'25.65"W-northeast;34˚23'28.65"S,53˚58'10.05"W-southwestand 34˚22'52.36"S,53˚35'12.89"W-southeast.It is bounded by the Atlantic Ocean on the south, by the Laguna Negra (Black Lagoon) on the east, by the Laguna de Castillos (Castillos Lagoon) on the west and by Sierra de La Blanqueada on the north.

Methodology for the Analysis of Thematic Information
A map produced by Zaffaroni et al. (2005) which states the palm density levels of the Castillos palm forest was used as base material.Five different density categories of the palm forest were determined by aerial photographic interpretation and field surveys (Figure 2).By using this methodology, an area of 11,611 hectares of palm forest was calculated and distributed according to the different density categories (Table 1).
The layers of digital information related to soil units and types, drainage, land   and rock formations, infrastructure and both cultivated and natural vegetation.
These are standard procedures for digital processing in land management (Filippini et al., 2012).These levels of information were included in the GIS environment by cropping the portion corresponding to the area of study.
The "palm forest" class of the land cover layer was replaced by the "palm forest density classes" layer (Zaffaroni et al., 2005) A vector layer of rural lots (basic administrative units in which the rural territory is organized) was overlapped (IDEuy, 2008).The area occupied by each class of land use in each lot was calculated using GIS.
As farms may be made up of one or more lots and this information is not officially digitalized in Uruguay, it was necessary to crosscheck the information on rural lots with registers from owners and tenants who manage cattle per lot from de Ministry of Farming, Agriculture and Fisheries.Registry units registered under the same name were grouped to form the different farms.This procedure allowed us to calculate the total area of each farm and of each land cover class in each farm.
During the springs of 2006 and 2007, seven surveys that lasted from 2 to 3 days were conducted in situ with the aim of mapping and geo-referencing vegetation units and other points of interest, using as support the 1:50,000 topographic maps and aerial photos at scale 1:20,000 obtained from the Military Geographic Service.Landscapes and plant communities were described by physiognomy.Samples of plant species of natural grasslands were obtained by using 50-meter transects every 50 centimeters.In the case of forests, wetlands and rocky outcrops, the most conspicuous species were recorded.Herbarium collections were made and photographs were taken for those cases in which it was ne-cessary to make laboratory observations and use keys for correct identification.
The correct names of the species were verified according to the Catalog of Vascular Plants of the Southern Cone (2016).150 sites were sampled and the average size was of about one hectare.Subsequently, using the list of plant species that had been identified, wild plant genetic resources in the landscape of the palm forest were determined by literature review., 1976).These units correspond to Typic Argiaquolls (49.03%)

Results and Discussion
and Typic Argiudolls associated to Pachic Vertic Argiudolls (32.37%), with poor to rather poor drainage.With regard to their use aptitude, they are comprised within the category of arable land under special conditions (57.76%) or non-arable land, suitable for the production of pastures and with strong limitations for forest use (27.01%).The distribution of the different density levels within the palm forest was not associated to soil types or use aptitude.

Plant Community Diversity, Species and Plant Genetic Resources
In the set of 150 surveyed sites on field trips we recorded a total of 302 plant taxa, distributed in 87 families and 224 genera (Table 3).As the identification was made only at the genus level in some cases, the number of taxa is not equiv-   alent to the number of species.It must be noted, however, that the goal was not to complete an exhaustive floristic list but to identify the most relevant species which are structuring the landscape.
Among the 302 taxa, we found 11 species that had been previously identified by Marchesi et al. (2013) as of high priority for conservation in Uruguay based on endemism, rarity and limited distribution criteria, in addition to butia palms, which appear on the list of endangered species due to their lack of regeneration (Table 3).In terms of plant genetic resources, species for medicinal (130), ornamental (221), fodder (34), food (8), timber (12) and fiber (10) use were included in the group of 302 taxa (among many other uses).Among plant genetic resources, 106 multipurpose species stand out.
Natural grasslands or "campos" have the highest percentage of occupation in the studied area (Table 2), just as it happens throughout the Uruguayan territory.The geographic location of Uruguay in a subtropical-temperate climate transition zone, the geomorphological and soil diversity and the confluence of different floras have originated a wide diversity of "campos" types (Rosengurtt, 1943;Millot et al., 1987;Lezama et al., 2011), this being a typical feature of the Pampa biome.The natural grasslands of the palm forests are a particular case due to their botanical composition, which differs from that of the surrounding fields (Rivas et al., 2014).Within this context of species and ecosystem diversity, plant genetic resources have evolved, mainly grasses and legumes used for forage, which coexist in a network with herbaceous species used mainly for medicinal and ornamental purposes.
Wetlands, which occupy nearly 13% of the territory of the palm forest, include both permanently flooded lands and lands which are only flooded at certain times of the year.Among the typical environments, we distinguished marshes with floating and submerged rooted plants, watergrass or gramales, which are herbaceous formations consisting mainly of Luziola peruviana, a variety of monospecific groupings, including those of Scirpus giganteus, Schoenoplectus californicus, Eryngium pandanifolium and E. eburneum, Typha domingensis, Erythrina crista-galli, Juncus acutus, Salix humboldtiana, Cephalanthus glabratus, Solanum glaucophyllum, and panicgrass or pajonales, which are dense communities of Panicum prionitis, Erianthus angustifolia or Paspalum quadrifarium (Alonso, 1998;Rivas, 2013).Acid wetlands are yet another particular case found in this landscape (Alonso & Bassagoda, 2006).Among the best known plant genetic resources in wetland environments, we distinguish those that are used in construction, furniture making, basketry and handicrafts.
The proportion of native forests in this landscape is 3 -4 times higher than the average for Uruguay of around 4% (Table 2).Different types of forests and scrublands were identified, which are named according to their topographical location: montane, coastal and riverine (Brussa & Grela 2007).Another very particular community called "ombu forest" (Phytolacca dioica) is found, being the largest grouping of this species in the country, which stretches along 20 kilometers by the shores of Laguna de Castillos (Castillos Lagoon) (Rodríguez-Gallego, 2006).Among the different plant genetic resources found in the forests, we may highlight those with food, timber, medicinal and ornamental use.
Within this context of ecosystem diversity, coastal communities (Fagúndez & Lezama, 2005) and vegetation accompanying rocky outcrops are also found.
Palm forests (Figure 3), as a specific type of forest, are mainly located on the middle plains, creating different landscapes not only due to their varying density levels, but also due to the differences in the accompanying vegetation and their associated anthropic uses (Rivas, 2013).Nowadays, butia palms are mainly used Figure 3. Butia odorata: Palm forest, palm and butia.
for food.Also, leaf fibers are used in handicrafts and the whole plant is used in landscaping projects with ornamental purposes.Sifted pulp, fiber and seeds of high nutritional value are obtained from fresh fruit products and used in the production of a wide range of foods such as ice cream, dessert sauce, jam and marmalade, sweet and sour sauce, stuffing for chocolates, cookies and pralines (Crosa et al., 2014), in addition to the typical and traditional butia liqueur.

Palm Forest Distribution in Agricultural Land
Within the surveyed area there were a total of 2153 lots, which added up to an area of 158,450.8hectares.This number is different to the total area of the surveyed territory (Table 2) as some rural lots are not entirely located within the surveyed area.
There are 700 rural lots where palm forests are found (32.5%), with varying proportions that range from less than 1% to 100% of the total area.Moreover, a relatively low proportion of these lots have a high percentage of palm forest coverage.A total of 117 lots have more than 50% of their area covered by palm forest that could be considered relevant to the conservation plan.However, in order to effectively determine the importance of these lots in the design of the conservation area, it is necessary to link this information to the area which is indeed occupied by each of them as well as to identify the farms and their location within the general context of distribution of the palm forest and its different density levels.
By comparing the information on lots and the database of owners and tenants who manage cattle we were able to sort into farms 77% of the cases (539/700 lots), a relatively satisfactory scenario considering they cover 87% of the area of the Castillos palm forest.
We managed to configure a total of 212 farms where the palm forest is found, 101 of which have all lots covered by palm forest and 111 of which have lots where the palm forest may or may not be found.Farms where palm forest is found have a size that ranges from 5.15 to 3848 hectares and a modal value of 78 hectares.The majority of them, more specifically 86.8%, have less than 500 hectares, a slightly higher proportion than the average for the department of Rocha (78.5%) (DIEA, 2013).These farms are considered units of family agriculture according to Uruguayan law, which states that farmers can not exploit an area of more than 500 hectares with CONEAT 100 (an index related to the average production capacity of the country) and they must live within 50 km of the property, have no more than two employees and that agriculture has to account for their main source of income (MGAP, 2009).The group of farmers with less than 20 hectares (11.3%) is also slightly above the average for the department of Rocha (9.5%).
Out of the 10,094.3hectares of land covered by palm forest, 9241 hectares are located in farms of more than 100 hectares and 7174.11hectares in farms of more than 500 hectares (Figure 4).The Pearson correlation coefficient between farm size and palm forest area is of 0.70.This result becomes a crucial element By overlapping layers of farms and palm forest we noticed that 18 farms (Figure 5), all of which are characterized by having more than 100 hectares of palm forest, have a total of 6952.40 hectares of palm forest (68.9% of the palm forest is located within farms) and account for 45.5% of the total area of the farms.This situation had not been previously quantified, so it becomes an important element in the planning and formulation of a management plan for palm forest conservation.These 18 producers own substantial areas of palm forest of very low density, while 6 producers own areas of more than 100 hectares of low density.The importance of these low and very low density palm forests is that, under monitored grazing conditions, regeneration can take place (Rivas & Barbieri 2015).In the palm forest area with the highest density, hardly any new butia plants can be found under conditions of exclusion or low stocking rates of grazing, probably due to the creation of an environment with greater humidity and less luminosity.Palm forests of medium density are also well represented by all these farmers.
Areas of palm forest with high and very high density levels occupy, in general terms, a relatively low area (Table 1) for this group of 18 properties.However, in a second group of farms where there are between 50 and 100 hectares of palm forest, two farms stand out for having areas of dense and very dense palm forest.
By including these two farms as part of the proposed conservation area, palm forests of all density levels are represented.It would also promote the development of ecotourism and the marketing of butia by-products, which are particularly found in this area of high aesthetic value and easy access known as "Vuelta del Palmar".A reserve comprising these 20 farmers is considered appropriate, taking into account the heterogeneity of the environments, the quality of the available habitat, a low edge effect and the concept of minimum viable population (Van Dyke, 2008;Hodgson et al., 2011).Although there are no studies that define the need for a certain population size of butia palms, it would be large enough so as to ensure the conservation of the palm forest.The connectivity of the area, which depends on the potential spread of the species and on habitat fragmentation (Luque et al., 2012), can be considered appropriate, although studies on the spread of pollen and seeds of butia palms are scarce (Rodríguez-Mazzini & Molina, 2000).Nor is there information on gene flow or on whether palm forests are a single population or they are structured in more subpopulations.The observed fragmentation is mainly explained by natural environmental conditions such as the presence of hills and wetlands, which are unsuitable environments for butia palms.

Conclusion
The main contributions of this work for the establishment of a conservation area within the butia palm forest of Castillos are the set up of a Geographic Information System (GIS) and the survey of plant communities, species and genetic resources.This GIS allowed us to determine the environmental conditions and the distribution of the palm forest and its density levels in each farm.By analyzing this information, we concluded there are 212 farms in which the palm forest is found, the larger ones having greater proportions of palm forest.The presence in the forest palm landscape of a high ecosystemic, specific and genetic richness values the territory for its conservation.Valuable plant resources with current and potential uses stand out among the 302 surveyed taxa as well as 11 species of high conservation priority apart from butia palms.
As this is an agricultural area that belongs to private owners, the focus of any proposal should be based on the coexistence of production practices with management techniques that can be adapted to conservation strategies and the inclusion of farmers and local dwellers in the definition and management of the area (Bridgewater, 2016).In this sense, the focus of Biosphere Reserves (PROBIDES, 1999;Pezzani, 2007;Ozyavuz & Yazgan, 2010), some categories of IUCN protected areas such as Protected Landscapes (IUCN, 2008), the Regional Natural Parks (SNAP, 2010) and the GIAHS (Globally Important Agricultural Heritage Systems) (Koohafkan & Cruz, 2011) are valid options to implement the conservation of Butia palm forests within the framework of integrated biocultural landscape conservation alternatives.Management plans and conservation policies should be framed within a context of sustainable agricultural systems (Dale et al., 2013;Benoit et al., 2012;Baiamonte et al., 2015).
are also grateful to the Federal University of Pelotas (Brazil), Embrapa Temperate Climate (Brazil) and CAPES (Brazilian Agency for the Improvement of Higher Education Personnel) for their valuable cooperation and contribution to the conservation and sustainable use of Butia.We especially thank Martin Jaurena, César Fagúndez, Maria Puppo, Inés Espasandín and Jerónimo Pardiñas for their support in field work, Patricia Mai for her collaboration in the organization of taxonomic information, Federico Condón, Laura del Puerto and Mariana Vilaró for their assistance in writing this paper and the farmers and organizations for their cooperation and permission to have access to their lands.
Figure 1.Map of South America, Uruguay and the Department of Rocha (left).Delimitation of the area of the Castillos palm forest (Rocha, Uruguay).

Figure 5 .
Figure 5. Map showing the location of the 18 priority properties to integrate the conservation area.

Table 2 .
Land cover categories and their respective areas in the palm forest territory of Castillos (Rocha, Uruguay).

Table 3 .
Plant native taxa, priority species for conservation and plant genetic resources surveyed in the palm forest territory of Castillos.