Study extent
The MIGRAME dataset covers the Pyrenean oak forests in Sierra Nevada mountain range.
Quercus pyrenaica forests
The Pyrenean oak (Quercus pyrenaica Willd.) forests extend through southwestern France and the Iberian Peninsula (Franco (1990)) (Figure 2a). In the Iberian Peninsula these forests live under meso-supramediterranean and mesotemperate areas and subhumid, humid and hyperhumid ombroclimate (Rivas-Martínez et al. (2002)) living on siliceous soils, or soils impoverished in basic ions (Vilches de la Serna (2014)). Q. pyrenaica requires between 650 and 1200 mm of annual precipitation and a summer minimal precipitation between 100 and 200 mm (García and Jiménez (2009), Martínez-Parras and Molero-Mesa (1982)), being the summer rainfall a key issue on the distribution of the specie (Río et al. (2007), Gavilán et al. (2007)).
The forests dominanted by Q. pyrenaica are an ecosystem included in the Annex I of the Habitat Directive (habitat code 9230: Quercus pyrenaica oak woods and Quercus robur and Quercus pyrenaica oak woods from Iberian northwestern). The conservation status of this habitat is not well known (EIONET (2014)), partly due to lack of detailed ecological studies (García et al. (2009)).
This species reaches its southernmost European limit at Sierra Nevada mountains, where eight oak patches (2400 Has) have been identified (Figure 2b), ranging between 1100 - 2000 m a.s.l. and generally associated to major river valleys. Sierra Nevada is considered a glacial refugia for deciduous Quercus species during glaciation (Brewer et al. 2002, FEM; Olalde et al 2002; Rodríguez-Sánchez et al 2010) and these populations are considered as a rear edge of the habitat distribution, which is important in determining habitat responses to expected climate change (Hampe and Petit (2005)).
These forests, as other vegetation types, have suffered intense pressure from human use (used for extracting wood, grazing, etc.) which has reduced their distribution area and in some cases their use has led to changes in their floristic pattern (Gavilán et al. (2000), Gavilán et al. (2007)).
Q. pyrenaica is considered as vulnerable in southern Spain (Vivero et al. (2000)). The populations of Pyrenean oak forests at Sierra Nevada are considered relict forests (Vivero et al. (2000), Melendo and Valle (2000)) and they have suffered an intensive anthropic use in the last decades (Camacho-Olmedo et al. (2002), Valbuena 2010). The relictic presence of this species in Sierra Nevada is related both to its high genetic resilience (Valbuena Carabana and Gil 2013, Tree Genetics & Genomes (2013)) and to its elevated intraspecific genetic diversity (Valbuena CArabana and Gil, Parques Nacioanles). However, they are also expected to suffer the impact of climate change, due to their climate requirements (humid summers). Thus, simulations of the climate change effects on this habitat pointed out a reduction of its suitable habitat for Sierra Nevada (Benito et al. (2011)).
Sampling description
We sampled at two localities of the Pyrenean oak forests at Sierra Nevada: Robledal de Cañar and Robledal de San Juan. We selected those two sites based on previous works (Pérez-Luque (2011), Pérez-Luque et al. (2013)) that clustered the populations of Q. pyreanica forests based on their plant species composition and environmental features. The Robledal de Cañar site (1366-1935 m a.s.l., 37°57'28.04''N, 3°25'57.1''W; Cáñar, Granada, SE Spain) was located in the Alpujarras Region on the southern slopes of Sierra Nevada. The Robledal de San Juan (1189-1899 m a.s.l., 37°7'29.63''N, 3°21'54.60''W; Güejar-Sierra, Granada, SE Spain) site was located in the northern slopes of Sierra Nevada.
The sampling desing was determined by the hypothesis of the project (see Project Design description section).
Altitudinal migration desing.
To test hypothesis of altitudinal migration, we sampled a total of 104 transects distributed along an altitudinal gradient in the two sites. We sampled two transects (separated at least 10 m from each other) every 25 m of elevation from forest limit to treeline ecotone at both study sites. In each locality we performed three replicates of this desing.
Habitat colonization desing.
To test the hypothesis of colonization of marginal habitats, we performed transects in two types of marginal habitats: abandoned agricultural areas and pine plantations. A total of 64 transects were located within the marginal habitat and on the edge between marginal habitat and pyrenean oak forest. The number of transects inside the marginal habitat was determined by the size of the marginal habitat.
Forests samplings.
In addition to the above surveys, we conducted a survey inside Q. pyrenaica forests. A total of
31 transects were distributed in the two sites.
Data collection
We sampled a total of 199 linear transects of 30 m x 10 m. Within each transect all tree species was mapped and the species identity annotated. Diameter size and tree height was measured for all individuals. duda diametro mayor. Field data were recorded using handheld PDAs. A customized application (app) was built to facilitate both data collection and store (Pérez-Pérez et al. (2013)). Automatic integration of the data into an information system was done thanks to this app.
Quality control
The transects coordinates were recorded with a handheld Garmin eTrex Vista Global Positioning System (GPS, +-5 m accuracy, Garmin (2007)) (ED1950 Datum). We also used colour digital orthophotographs provided by the Andalusian Cartography Institute and GIS (ArcGIS 9.2; ESRI, Redlands, California, USA) to verify that the geographical coordinates of each sampling plot were correct (Chapman and Wieczorek 2006).
The specimens were taxonomically identified using Flora Iberica (Castroviejo et al. 1986-2005, Castroviejo 2001). The scientific names were checked with databases of International Plant Names Index (IPNI 2013) and Catalogue of Life/Species 2000 (Roskov et al. 2013). We also used the R packages taxize (Chamberlian and Szocs 2013, Chamberlain et al. 2014) and Taxostand (Cayuela and Oksanen 2014) to verify the taxonomical classification.
We also performed validation procedures (Chapman 2005a, 2005b) (geopraphic coordinate format, coordinates within country/provincial boundaries, absence of ASCII anomalous characters in the dataset) with DARWIN_TEST (v3.2) software (Ortega-Maqueda and Pando 2008).
Method Steps
- All data were stored in a normalized database and incorporated into the Information System of Sierra Nevada Global-Change Observatory (http://obsnev.es/linaria.html – Pérez-Pérez et al. 2012; Free access upon registration). Taxonomic and spatial validations were made on this database (see Quality-control description). A custom-made SQL view of the database was performed to gather occurrence data and other variables associated with some occurrence data (diameter size and tree height of each individual).
The occurrence and measurement data were accommodated to fulfil the Darwin Core Standard (Wieczorek et al. 2009, 2012). We used Darwin Core Archive Validator tool (http://tools.gbif.org/dwca-validator/) to check whether the dataset meets Darwin Core specifications. The Integrated Publishing Toolkit (IPT v2.0.5) (Robertson et al. 2014) of the Spanish node of the Global Biodiversity Information Facility (GBIF) (http://www.gbif.es:8080/ipt) was used both to upload the Darwin Core Archive and to fill out the metadata.
