Proyecto AVIS: a Spanish open access bird database available for research

. Proyecto AVIS 1 is an open access citizen science database that stores information collected by amateur ornithologists about bird occurrences and abundance in Spain. Proyecto AVIS was launched in 2005 and today stores data from 415 species ( ca. 90% of bird species in Spain); it covers 30% of the Spanish territory, including the Canary Islands in the Atlantic Ocean and the Balearic Islands in the Medi-terranean Sea. Here, we acknowledge the work of all the volunteers that have gathered bird records in the field and uploaded these observations over the last 10 years, and introduce Proyecto AVIS to a broader community of biogeographers and macroecologists to promote its use for research.


Structure of the database
The database is programmed in MySQL, an opensource relational database management system. Proyecto AVIS has two groups of tables: data tables and dictionary tables. Data tables compile the names of observers, bird observations, observation periods per day and observer effort per period. Dictionary tables show standardized information per field: species, age, sex, UTM coordinates, Spanish region, reproductive status, phenological category, habitat, altitude and atmospheric conditions (visibility, wind speed, rainfall, etc.). The robustness of the data stored in Proyecto AVIS has been promoted by the standardization of its fields. Users select the species, habitat and other main variables from a list of pre-defined categories. Thus, spelling mistakes and synonyms (e.g. species or habitat synonyms) are avoided. The Proyecto AVIS database only allows free text in the 'Notes' field, where extra information can be added.
Users must enter bird records using the 'New observation' formulary, which consists of several fields. The first seven fields are mandatory and the others are optional. Below we describe the fields, indicating which are text and which are ISSN 1948-6596

Sampling effort
The 87,000+ observations in Proyecto AVIS cover 1750 different 10x10 km UTM squares, representing 30% of the total area of Spain (Fig. 1). Unfortunately, UTM coverage is geographically uneven: there are regions with few records (e.g. Galicia, in the northwest of the Iberian Peninsula) while other areas have almost 100% coverage (e.g. Mallorca). We hope that the currently unsampled UTM squares will receive increasing attention in the near future. We are also now working to extend map coverage to North Africa, to obtain data from the Strait of Gibraltar, an essential migration route between Africa and Europe where some bird conservation societies work, such as COCN 5 . Some of the observations include information about the sampling effort of the volunteers (hours in the field). At the time of writing this paper, Proyecto AVIS contained such information for 64 transects covered by foot, 82 transects covered by car and 778 stationary observations. On average, observers spent approximately 4 hours in the field per survey, with maximum values exceeding 11 hours (Table 1). Transects by car allowed de-tection of more species than other methods, and stationary observations the fewest ( Table 1).

The web interface
The Proyecto AVIS web page 1 allows queries to explore the database and obtain general information and statistics about observations. In the 'Maps and Queries' section, (Mapas y Consultas) any user (registered or not) can find tables, maps and graphs from observations included in the database. Below, we describe different query possibilities. 1. Advanced queries: in this section users can obtain observations by filtering using different fields, and can display the results as a list of observations or by grouping them by different criteria such as species, date, regions or UTM squares. Moreover, it is possible to plot graphs and distribution maps using UTM squares with the presence of the selected species. Registered users can also download the information in data files (.csv). Registration is simple and free. 2. Observation periods: users can view periods, select one and get the number of observations by species, the list of contributors as well as all the general information about the period (weather, duration, etc.

rAvis R-package
We have recently developed rAvis 7 , an Rpackage to download and visualize the data from Proyecto AVIS in the R-environment (Varela et al. 2014). The rAvis package is part of the rOpenSci project 8 , which enables scientists to easily use the data stored in open access databases for research.

Potential uses of Proyecto AVIS
In spite of being one of the oldest citizen science web projects (started in 2005), macroecologists and biogeographers are mostly not aware of the existence of Proyecto AVIS. We would therefore like to conclude the presentation of Proyecto AVIS by encouraging researchers to use this large bird database.  Potential applications are very wide. The database can be explored using the rAvis Rpackage and the downloaded information can be combined with the researcher's own data or with other online bird databases in order to complete the species samples (e.g. eBird 9 or GBIF 10 ). Occurrence data can be used to calibrate ecological niche models and project them into different climatic scenarios to map species' distributions in the past or in the future. Detailed information about particular species, combined with GIS layers, could help researchers to understand the impacts of the land-use change in Spain. Further, seasonal information about species' occurrences could be used to analyze changes in species diversity during the year, or information from several years could be used to assess long-term changes in bird species diversity. Finally, conservation strategies could be tested with these temporal series of species' occurrences.