(C) 2013 Dan Cogălniceanu. This is an open access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC-BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
For reference, use of the paginated PDF or printed version of this article is recommended.
Nineteen species of amphibians inhabit Romania, 9 of which reach their range limit on this territory. Based on published occurrence reports, museum collections and our own data we compiled a national database of amphibian occurrences. We georeferenced 26779 amphibian species occurrences, and performed an analysis of their spatial patterns, checking for hotspots and patterns of species richness. The results of spatial statistic analyses supported the idea of a biased sampling for Romania, with clear hotspots of increased sampling efforts. The sampling effort is biased towards species with high detectability, protected areas, and large cities. Future sampling efforts should be focused mostly on species with a high rarity score in order to accurately map their range. Our results are an important step in achieving the long-term goals of increasing the efficiency of conservation efforts and evaluating the species range shifts under climate change scenarios.
Amphibia, species distribution, species range, biodiversity data, species richness, rarity
Biodiversity studies have intensified after the Convention on Biological Diversity was signed at the 1992 World Summit in Rio de Janeiro. Despite repeated attempts to halt biodiversity loss, the 2010 targets have not been met (
Mapping species distribution data is a popular and successful way of assembling and presenting spatial information on a variety of organisms, including amphibians (
Eastern European countries do not usually provide quality distribution data due to less uniform and intensive recording effort (
We used occurrence records from four major sources: published data, museum collections, personal communications from specialists, and our own unpublished field data. The records were stored and managed in a Microsoft Access database and later imported in a GIS environment as geodatabase. For spatial representation we used the Universal Transverse Mercator (UTM) grid system which is the most frequently used cartographic framework for mapping species distributions (
The distribution records with a spatial resolution ≤ 25 km2 were georeferenced and assigned the corresponding alphanumeric UTM 5 × 5 km grid cell code. The occurrences with a spatial resolution > 25 km2 were assigned in Google Earth (Google Earth v7.0.2, Google Inc., CA) to a single grid cell, based on expert knowledge of the species’ habitat requirements (
Several records were not used for mapping species distributions when they could not be referenced to a specific locality or toponymy (e.g., distribution records assigned to mountain ranges, geographical provinces or hydrographic basins, without a finer scale reference), or unspecified taxa within a genera, while other records were doubtful or erroneous (for a detailed description of errors see
Three pairs of amphibian species hybridize in Romania: Bombina bombina with Bombina variegata, Lissotriton vulgaris with Lissotriton montandoni, and Triturus cristatus with Triturus dobrogicus. In most cases, occurrences were assigned to a hybrid based on the analysis of morphological features. The grid cells in which hybrids occurred were mapped distinctly for both parental species. The data used for mapping the hybrid zones varied in quality: while some reports refer to “hybrids” or “intergrades”, others were identified based on reports of both parental species occurring in the same grid cell. Since the large water frogs of the Pelophylax complex (esculentus and ridibundus) are difficult to distinguish (
To identify the potential bias in sampling effort, we first counted the number of amphibian records per grid cell. Then, we used Global Moran’s I to assess the general trend of spatial autocorrelation in occurrences across the entire country. If the Moran’s I test is statistically significant the spatial pattern of amphibian records per grid cell can be spatially clustered (Z > 0), dispersed (Z < 0) or random (Z = 0) (
The mean altitude of 5 × 5 km grid cells was extracted from SRTM data (
We calculated the Extent of Occurrence (EOO) as a 100% minimum convex polygon using the state boundary as a mask, in order to remove the areas outside Romania, where no sampling was accounted. We estimated the Area of Occupancy (AOO) as the total area of 5 × 5 km grid cells where a species was reported (
To map species richness we aggregated the data to a spatial resolution of 50 × 50 km UTM grid cells. Mapping the species richness at a coarser resolution reduced the potential bias in sampling effort and allowed a better understanding and visualization of regional patterns (
All spatial analyses were performed in ArcGISDesktop 10 (ESRI, CA), with α = 0.05.
In total, 26779 amphibian occurrences were compiled and stored in a geodatabase. We collected 20868 records from 201 published papers (Appendix 1). Most papers (59%) were published after 2000 (Fig. 1). We compiled further 5381 unpublished records from our own field surveys and 530 museum collection records. The majority of occurrences were dated after 1990 (89%) and only 11% represented old records collected before 1990. The detailed occurrence statistics for each species are presented in Table 1. Our database increased the average number of amphibian records per 100 km² compared with GBIF dataset (http://data.gbif.org, accessed 15.02.2013) from 0.123 in the GBIF dataset to 11.2 in our database (Table 2).
Number of publications containing amphibians’ distribution data (1954–2012).
The occurrences and measures of amphibian species range in Romania. Extent of Occurrence (EOO) was estimated as 100% minimum convex polygon, and Area of Occupancy (AOO) as the total area of 5 × 5 km UTM cell containing species records. Since not all grid cells matched the 25 km² area, the computed AOO is not a multiple of this value.
Species | Total number of records |
New records (after 1990) |
Total number of UTM5 cells |
EOO (km2) |
AOO (km2) |
Rarity index |
---|---|---|---|---|---|---|
Salamandra salamandra | 1200 | 1033 | 775 | 132910 | 29769 | 50.5 |
Triturus cristatus | 1639 | 1404 | 436 | 190721 | 40763 | 31.7 |
Triturus dobrogicus | 209 | 180 | 145 | 189130 | 4974 | 75.6 |
Ichtyosaura alpestris | 768 | 629 | 962 | 108965 | 19161 | 59.3 |
Lissotriton vulgaris | 2114 | 1846 | 1220 | 228999 | 52377 | 17.1 |
Lissotriton montandoni | 569 | 429 | 284 | 56773 | 14163 | 78.9 |
Bombina bombina | 1720 | 1569 | 936 | 235347 | 42033 | 21.9 |
Bombina variegata | 3116 | 2811 | 1445 | 151826 | 77580 | 43.1 |
Pelobates fuscus | 618 | 518 | 399 | 230794 | 14957 | 40.7 |
Pelobates syriacus | 153 | 122 | 59 | 48292 | 3546 | 87.9 |
Bufo bufo | 2128 | 1968 | 1291 | 223802 | 52561 | 24.4 |
Bufo viridis | 2006 | 1791 | 1299 | 235801 | 48864 | 12.2 |
Hyla arborea | 1801 | 1676 | 1156 | 233887 | 44265 | 11.4 |
Rana dalmatina | 2027 | 1833 | 1270 | 225426 | 49982 | 25.2 |
Rana arvalis | 351 | 270 | 215 | 88423 | 8638 | 69.1 |
Rana temporaria | 1806 | 1614 | 1020 | 132586 | 44930 | 49.6 |
Pelophylax lessonae | 203 | 195 | 108 | 190870 | 4867 | 73.2 |
Pelophylax kl. esculentus | 4351 | 3981 | 1928 | 236441 | 105855 | 5.7 |
Total | 26779 | 23869 | - | - | - | - |
The bias in amphibian species occurrence data for several European countries extracted from GBIF (downloaded on January 30, 2013).
Country | GBIF records | Country area (km2) | Average number of records per 100 km2 |
---|---|---|---|
Luxembourg | 8119 | 2666.7 | 304.461 |
United Kingdom | 101305 | 260565.6 | 38.878 |
Spain | 37764 | 544571.2 | 6.934 |
France | 24928 | 570427.7 | 4.370 |
Poland | 6753 | 314347.5 | 2.148 |
Hungary | 160 | 93374.6 | 0.171 |
Romania | 293 | 237318.9 | 0.123 |
Ukraine | 737 | 600856.2 | 0.122 |
Turkey | 452 | 794961.2 | 0.056 |
Greece | 61 | 133552.6 | 0.045 |
From a total of 9977 UTM 5 × 5 km grid cells covering the Romanian territory, 3013 grid cells (i.e., 30.1%) contain amphibian sightings (Fig. 2). The number of records per cell is highly skewed (skewness = 31.25). Global Moran’s I test revealed a clustered pattern in the number of amphibian records per cell (Z = 66.91, p < 0.001), thus suggesting a strong bias in nationwide sampling effort. Since 73% of the country surface had ≤ 2 records per cell, the Getis Ord Gi* spatial statistic did not identify the coldspots of sampling efforts as being statistically significant, at the specified threshold distance of 7100 m, but it effectively identified the hotspots of sampling effort (Fig. 3). Among these hotspots, the Măcin Mountains National Park had the highest sampling effort (mean Z = 24.36). Other hotspots of sampling effort were the Jiului Gorge National Park (mean Z = 2.47) and the Iron Gates Natural Park (mean Z = 1.68). There was an obvious trend of clustering the number of records in close proximity to cities (e.g., Iaşi, Piatra Neamţ, Oradea and Satu Mare) (see Fig. 3).
Amphibian occurrences in Romania. Records reported before 1990 were plotted as old records whereas those reported after 1990 were considered new records.
Hotspots of sampling efforts within Romania. The p value is < 0.05 when Z scores take values between 1.65 and 56.44, suggesting a highly clustered pattern in the number of amphibian occurrences per UTM 5 × 5 grid cell.
The species richness represented on the 50 × 50 km grid, excluding the incomplete grid cells, ranged from 3 to 16 (Fig. 4). Individual species (i.e., excluding Pelophylax kl. esculentus) had different distribution patterns with the Extent of Occurrence ranging from 48292 km2 (Pelobates syriacus) to 235801 km2 (Bufo viridis), the Area of Occupancy ranging from 3546 km2 (Pelobates syriacus) to 77580 km2 (Bombina variegata), and the rarity index ranging from 75.6 (Triturus dobrogiscus, the rarest species) to 11.4 (Hyla arborea, the widespread species) (see Table 1).
Amphibian species richness within Romania at a 50 × 50 km grid resolution.
Taking into account the mean altitude of 5 × 5 km occupied grid cells, the altitudinal gradient of amphibians varied between 0 and 2007 m, with Ichtyosaura alpestris and Rana temporaria occurring over 2000 m (Fig. 5). The amphibian species occurrence maps in Romania are presented in Figs 6–23. Several species are widespread within the entire country (e.g., Lissotriton vulgaris, Bufo viridis) while others are restricted only to higher slopes (e.g., Lissotriton montandoni, Ichtyosaura alpestris, Salamandra salamandra), to plains (e.g., Pelobates syriacus) or floodplains (e.g., Triturus dobrogicus).
Altitudinal distribution of amphibian species in Romania (dot – mean, horizontal line – median, vertical bar – range, Ssa – Salamandra salamandra, Tcr – Triturus cristatus, Tdo – Triturus dobrogicus, Ial – Ichtyosaura alpestris, Lvu – Lissotriton vulgaris, Lmo – Lissotriton montandoni, Bob – Bombina bombina, Bov – Bombina variegata, Pef – Pelobates fuscus, Pes – Pelobates syriacus, Bub – Bufo bufo, Buv – Bufo viridis, Hya – Hyla arborea, Rda – Rana dalmatina, Rar – Rana arvalis, Rte – Rana temporaria, Ple – Pelophylax lessonae, Pke – Pelophylax kl. esculentus).
Salamandra salamandra.
Triturus cristatus.
Triturus dobrogicus.
Ichtyosaura alpestris.
Lissotriton vulgaris.
Lissotriton montandoni.
Bombina bombina.
Bombina variegata.
Pelobates fuscus.
Pelobates syriacus.
Bufo bufo.
Bufo viridis.
Hyla arborea.
Rana dalmatina.
Rana arvalis.
Rana temporaria.
Pelophylax lessonae.
Pelophylax kl. e sculentus and ridibundus.
By compiling the current database of amphibian distribution in Romania, we made available comprehensive distribution maps for further analyses, and provided the first nationwide spatial statistical analysis of the amphibians’ occurrence records.
The number of occurrences varies among species, the largest number of records (i.e., 16%) being that of Pelophylax kl. esculentus, a taxon including two species (Pelophylax esculentus and Pelophylax ridibundus). The altitudinal range of Pelophylax kl. esculentus varies between 0 and 1255 m. The highest number of single taxa occurrences belongs to Bombina variegata (i.e., 12%), a species with a broad altitudinal gradient (80-1988 m). Despite the large number of occurrences, Bombina variegata is not the most common species in Romania (rarity index = 43.1), yet it has a high detectability (i.e., easy to recognize, is active during the day and occurs mostly on the road-side temporary ponds (
The number of occurrences per species is negatively correlated with the rarity index (Spearman rho = -0.8, p < 0.001), indicating that the range of many species is scarcely known. For some species, the large number of occurrences is not reflected in a high rarity score. However, the ranges of Bombina variegata and Rana temporaria seem to be accurate since they have a high number of occurrences, but a moderate rarity score. At the species level, the future sampling effort should be focused mostly on species with a high rarity score in order to accurately map their range.
A general feature of a research study is that data are considered as a means to an end (i.e., publication), and are therefore treated as consumables (
The incomplete and biased species inventory in Romania may have several causes: difficult access due to low road density, complex landscape (with 15% of the territory above 800 m), limited funds available for large-scale inventory and monitoring projects, and lack of institutional support. For instance, no species distribution databases are publicly available at the Romanian Ministry of the Environment (
The low quality of biodiversity datasets and the lack of standards in publication and data management limit their usefulness (
We are grateful to the following persons for sharing their distribution data with us: Dr Arntzen Jan, Dr Bănărescu Petru, Dr Bereş Iosif, Buhaciuc Elena, Dr Gâldean Nicolae, Dr Hartel Tibor, Dr Kyek Martin, Dr Pârvulescu Lucian, Dr Oţel Vasile, Dr Skolka Marius, Sós Tibor, Talbot Neil, and Tallowin Oliver. We thank Dr Franco Andreone, Dr Antonio Romano, and an anonymous reviewer for their constructive comments on the paper. This work was supported by two grants of the Romanian National Authority for Scientific Research, CNCS-UEFISCDI, project number PN-II-RU-TE-2011-3-0183 (principal investigator Laurenţiu Rozylowicz) and CNCS-UEFISCDI, project number PN-II-ID-PCE-2011-3-0173 (principal investigator Dan Cogălniceanu).
Publications used to compile distribution of amphibian species native to Romania. (doi: 10.3897/zookeys.296.4872.app) File format: Microsoft Word document (doc).