Exploring the geographic distribution of migratory birds in Argentina: first experience using recently collected citizen-science data and ecological niche modeling

DI GIACOMO, ADRIÁN; RABUFFETTI, FABIAN; URSINO, CYNTHIA A.; VOLPE, NOELIA

2009-08-01/2010-12-31

Biológica

Prom.Gral.del Conoc.-Varias ciencias

Migratory shorebirds are vulnerable to natural and anthropogenic threatsthat affect them during migration and in wintering grounds. Updated informationabout geographic distribution of shorebirds during their migration or winteringseasons is key for conservation actions. In Argentina there are 22 species ofmigratory shorebirds from the Northern Hemisphere and the available knowledgeon their geographic distribution is mainly based on the maps that appear on birdidentification guides or a few specific articles. We evaluated the effectiveness ofonline collection of biological data obtained by amateur birdwatchers to estimatethe geographic range of Northern Hemisphere migratory shorebirds wintering inArgentina in a ´citizen-science´ project. Through advertisement in social networks tojoin the eBird initiative in Argentina (http://ebird/content/Argentina) we collectedrecords of 20 migratory species during the spring and summer 2013. We combineddata from these bird-occurrence records with bioclimatic and environmentalvariables using a niche-modeling tool that allowed us to obtain more accuratedistribution maps. We used the algorithm of maximum entropy (MAXENT). Themodels obtained fitted very well to the final set of dependent variables that wereselected during the analysis, in all cases reaching very high values of the statisticalestimator (AUC> 0.750). To evaluate the effectiveness of these models, that wereobtained with a very low sampling effort and during a very short period of time,(´citizen science 2013? dataset), we compared such maps with published maps andother maps of potential distribution assembled using the same bioclimatic andenvironmental variables, but training models with two different datasets: 1) a set ofdata obtained from literature and museum specimens (´research-science? dataset),and 2) a set of data collected over 12 years in the global database of ebird.org(´citizen-science 2002-2013? dataset). The comparison of maps obtained withdifferent datasets indicates that the data gathered from birdwatchers during asingle season, in combination with ecological niche modeling, allowed us to obtaingood estimates of the geographic distribution of species of migratory shorebirds. Inparticular, the distribution maps of species associated with grasslands of centraland northern Argentina obtained through our one-season sampling and the morecomplete dataset resulted very similar. However, for species with larger ranges orwith ranges covering several ecoregions, the models obtained with the differentdatasets under comparison were less similar . Such differences are probablyassociated with different sampling effort or lack of data along the geographicalrange or ecoregions. For improving the effectiveness of annual estimates ofdistributional ranges using birdwatcher data, we recommend organizing volunteersto cover all ecoregions, as much as possible, under similar efforts. This could beachieved, for example, by designating the best areas to be covered bybirdwatchers nucleated in the 78 "birders clubs" (?COAs?) found in 23 of the 24provinces of Argentina. Finally, to balance the sampling effort we recommend anew design based on a grid, and the setting of a minimum sampling effort to beconducted in each ecoregion (i.e., a number of surveyed quadrates), similar to thetraditional ornithological "atlas", but incorporating eBird for collecting andprocessing data.