Introduction
Following a dramatic decline in both range and population in the second half of the 20th century, the Eurasian otter Lutra lutra is recovering in several European countries, including the UK (Crawford, Reference Crawford2003), Sweden (Roos & Rigét, Reference Roos, Rigét, Prigioni, Loy, Balestrieri and Remonti2011), Denmark (Elmeros et al., Reference Elmeros, Hammershøj, Madsen and Søgaard2006), France (Janssens et al., Reference Janssens, Defourny, De Kermabon and Baret2006), Spain (Ruiz-Olmo & Delibes, Reference Ruiz-Olmo and Delibes1998) and Italy (Prigioni et al., Reference Prigioni, Balestrieri and Remonti2007). This recovery is probably largely attributable to efforts in the 1970s to ban persistent organic pollutants, especially synthetic pesticides (lindane, DDT) and chlorinated biphenyls (Ruiz-Olmo et al., Reference Ruiz-Olmo, Lafontaine, Prigioni, López-Martín and Santos-Reis2000).
Although most industrially developed countries have made progress in controlling pollution and implementing conservation measures, in transitional or developing economies biodiversity governance is hindered by ongoing institutional change and the need to promote economic development, which often result in overexploitation of natural resources (CEC, 2001; Kluvánková-Oravská et al., Reference Kluvánková-Oravská, Chobotová, Banaszak, Slavikova and Trifunovova2009). Approximately 60% of the countries within the Eurasian otter's range have been listed as emerging and developing economies by the International Monetary Fund (IMF, 2012), and the species’ population trend in the eastern part of its range was the basis for its categorization as Near Threatened on the IUCN Red List in 2004 (Ruiz-Olmo et al., Reference Ruiz-Olmo, Loy, Cianfrani, Yoxon, Yoxon and de Silva2008).
Information about the current status of the species in many developing countries is scarce. Although the otter population is increasing in Poland (Romanowski, Reference Romanowski2006) and may not have declined in Hungary during the second half of the 20th century (Heltai et al., Reference Heltai, Bauer-Haáz, Lehoczki and Lanszki2012), the species has reportedly declined in Turkey (Kayaöz, Reference Kayaöz2002; Özdemir, Reference Özdemir2002), the South Caucasus region (Georgia, Armenia and Azerbaijan; Gorgadze, Reference Gorgadze2011), Belarus (Sidorovich, Reference Sidorovich2011), Nepal (Bhandari, Reference Bhandari2011) and Morocco (Delibes et al., Reference Delibes, Calzada, Clavero, Fernández, Gutiérrez-Expósito, Revilla and Román2012).
In Albania the communist era ended in 1991. The transition from centralized management of land and natural resources to a market-oriented economy has caused widespread damage to the environment (Jansen et al., Reference Jansen, Carrai, Morandini, Cerutti and Spisni2006), especially to aquatic ecosystems of the western lowlands, as a result of increased urbanization, overexploitation of woodland and ineffective waste management (Cullaj et al., Reference Cullaj, Hasko, Miho, Schanz, Brandl and Bachofen2005; Keukelaar et al., Reference Keukelaar, de Goffau, Pradhan, Sutmuller, Misurovic and Ivanovic2006).
Except for a few more recent surveys in the southern part of the country (rivers Seman and Drinos; Bego et al., Reference Bego, Malltezi, Beqiraj and Xhulaj2011; Hysaj et al., Reference Hysaj, Bego, Prigioni and Balestrieri2013), information about the distribution of the otter in Albania dates back to 1985, when during 18–27 May Prigioni et al., surveyed 31 sites, mostly in the western lowlands and in the upper catchment of the River Shkumbin (Prigioni et al., Reference Prigioni, Bogliani and Barbieri1986). The rationale for the survey method was based on Erlinge's recommendations; stretches of river up to 600 m long were surveyed for otter spraints and footprints on both banks (Chanin, Reference Chanin2003). A minimum distance of 200 m was surveyed per site and the otter was presumed absent if no sign was detected along the transect. To update information on the distribution of the Eurasian otter in Albania and to assess the impact of political and economic changes on the species’ conservation by comparing current and past range, we resurveyed the major watercourses and coastal lagoons of Albania in summer 2013.
Methods
Following the same protocol used by Prigioni et al. (Reference Prigioni, Bogliani and Barbieri1986), during 24 June–18 July 2013 we resurveyed the 31 sites surveyed in 1985, and an additional 42 sites, thus surveying all the major river catchments of Albania (Fig. 1). We used a 1 : 200,000 road map to select sites, which were chosen for ease of access, mainly at road bridges, to maximize the probability of finding signs quickly (Chanin, Reference Chanin2003). Sites were separated by at least 5 km. Twenty-nine sites were on main rivers, 26 on tributaries, seven on canals, five on lakes (Shkodrës, Ohrit, Vaut të Dejës, Fierzes) and six on coastal lagoons (Butrinti, Narta, Ceke, Merxhani, Patok, Viluni; Peja et al., Reference Peja, Vaso, Miho, Rakaj and Crivelli1996). Mean transect length was 587 m (range 380–1000 m). The most thoroughly surveyed catchments were those of the rivers Shkumbin (n = 13), Vjosa (n = 18), Mat (n = 6) and Ishëm-Erzen (n = 7); relative to their total length, few sites were located in the catchments of the rivers where otter occurrence had been confirmed recently. All sites were georeferenced using a global positioning system. To avoid bias as a result of variation in experience with sampling, all surveys were carried out by the same trained staff.
Sprainting sites were identified as places with spraints separated by at least 1 m from other spraints (Kruuk et al., Reference Kruuk, Conroy, Glimmerveen and Ouwerkerk1986). Following Prigioni et al. (Reference Prigioni, Bogliani and Barbieri1986) an index of otter marking intensity was calculated as the number of sprainting sites or spraints per 200 m. We assumed marking intensity represented a broad index of relative abundance (Mason & Macdonald, Reference Mason and Macdonald2004). The percentage of sites where otters were present in 1985 was compared to that found in 2013 (χ2 test), with respect to both the total number of sites (n = 73) and the number of sites surveyed in both years (n = 31). For the latter sample the distribution of sites with presence and absence was compared using a 2 × 2 summary table of coincidences and discrepancies.
For each site on a watercourse five habitat variables of potential importance to otters (Prenda et al., Reference Prenda, López-Nieves and Bravo2001; Remonti et al., Reference Remonti, Prigioni, Balestrieri, Sgrosso and Priore2008) were recorded visually in the field (Supplementary Table S1). Watercourse width and depth were estimated at intervals of c. 200 m and the mean values for each site were assigned to one of four classes (≤ 2.0, 2.1–6.0, 6.1–15.0 or > 15.0 m, and ≤ 0.5, 0.51–1.0, 1.1–2.0 or > 2.0 m, respectively). Water speed was categorized as stagnant, slow or moderate–high. The dominant land use within a 100 m wide belt on both river banks was estimated and categorized as bare soil, crops, pasture/shrubs or woods. Water pollution was categorized as absent (transparent water in ≥ 70% of watercourse depth, without any sign of pollution), low (less transparent water, solid waste on the banks), medium–high (clear signs of pollution, such as bad smell, turbidity, foam). Mean altitude was recorded for each site (≤ 250, 251–500, 501–750, > 750 m). To assess the influence of human disturbance on otter occurrence, using GE-Path v. 1.4.6 (Sgrillo, Ilhéus, Brazil), we assessed the overall percentage cover of agricultural and urban areas in a 4 km2 area centred on each site, by superimposing a 200 × 200 m grid on satellite images provided by Google Earth (Google, Mountain View, USA). Results were categorized as ≤ 25, 25.1–50, 50.1–75 or > 75%.
Otter presence/absence was then modelled against the predictor variables by backward stepwise logistic regression, using Wald's test to examine the statistical significance of each coefficient in the model. To avoid collinearity, which can inflate the standard errors of the estimates of model coefficients and produce unreliable results (Hosmer & Lemeshow, Reference Hosmer and Lemeshow1989), the correlation between the predictors was measured by Spearman's coefficient. Model variables were then selected, omitting those that represented redundant information; when two variables were correlated (P < 0.01), the one to be rejected was chosen according to the strength of its correlation with the dependent variable. Analyses were performed only for sites on watercourses, as most habitat variables could not be assessed precisely for lakes and lagoons.
Finally, from the last human population census (INSTAT, 2011) we obtained the population density (inhabitants per km2) of the counties (qarku; n = 12) and municipalities where sampling sites were located. For sites in the same administrative area the results were pooled to avoid pseudoreplication (Hurlbert, Reference Hurlbert1984), and mean values for positive and negative sites were compared using the Mann–Whitney test. The relationship between marking intensity and either altitude or population density was examined using Spearman's correlation test.
Results
Otter signs were detected at 44 (60.3%) of the 73 survey sites. From north to south, the species was widespread in the catchments of the rivers Drin, Mat, Shkumbin and Seman (where the otter occurred in the plain of Coriza) and absent on the rivers Gjanica and Vjosa. Otter signs were not found in the downstream stretches of the rivers Shkumbin, Seman and Vjosa, which cross the intensively populated plain facing the Adriatic Sea. No otter sign was found in the catchments of the rivers Ishëm (which runs through the counties of Durazzo) or Tirana and Bistrice, in the south, whereas otter presence was detected at one of four sites in the catchment of the River Erzen (Fig. 1). Otters were present in five of six lagoons and in the Lake of Ocrida (Ohrit), and absent from the Lake of Scutari (Shkodrës) and two of three artificial lakes (Liqueni Fierzes and Liqeni i Vaut të Dejës).
The percentage of surveyed sites where otters were present did not differ from that recorded in the 1985 survey (60.3 vs 54.8%; χ2 = 0.27, P = 0.61), even considering only the sites surveyed in both years (61.4 vs 54.8%; χ2 = 0.26, P = 0.61). The distribution of presence and absence was also similar: otter status in 2013 was the same as in 1985 for 71% of sites (χ2 = 10.9, P = 0.001). In contrast, marking intensity was lower in 2013 than in 1985 (Table 1).
Logistic regression identified percentage cover as the main variable affecting the occurrence of the otter (Table 2). Pecentage cover was negatively correlated with mean altitude (ρ = −0.42, P = 0.0006, n = 62) and positively correlated with water pollution (ρ = 0.44, P = 0.0003, n = 62).
1 % cover of agricultural and urban areas in a 4 km2 area centred on each survey site
2 Logistic coefficient
3 Odd ratios
Marking intensity was not correlated with any of the variables tested. Mean human population density of municipalities was significantly lower for sites where otters were present than sites where they were absent (62.3 vs 250.5 inhabitants per km2; U = 311, P = 0.028).
Discussion
The distribution of the otter in Albania in 2013 was similar to that recorded by Prigioni et al. (Reference Prigioni, Bogliani and Barbieri1986) in 1985. Otters were absent on the coastal stretches of the rivers Gjanica, Seman and Vjosa, where oil extraction and refinement remain the primary source of water pollution (Cullaj et al., Reference Cullaj, Hasko, Miho, Schanz, Brandl and Bachofen2005), whereas urban and industrial waste, together with increased urbanization, probably prevent colonization of the catchments of the rivers Ishëm and Erzen. The largest difference was recorded in the River Bistrice, the river with the highest marking intensity in 1985, where otters are now absent from two of three sites where they were present in 1985. In a report by the Government of Albania, direct killing has been indicated as the main threat to otters in the Southern Coastal Plain, but the River Bistrice has been canalized for much of its length and the whole watershed area has been degraded by uncontrolled urbanization (PAP/RAC-SOGREAH, 2005), suggesting that habitat alteration may have been a significant factor in the decline of the otter.
Although no habitat variable seemed to affect distribution at a local scale, confirming that otters are tolerant of a wide range of habitat conditions (Chanin, Reference Chanin2003), at the landscape scale the effect of land use and human population density may be regarded as an index of the impact of human activities on otter occurrence. Similarly, on the Iberian Peninsula otter distribution was reported to be affected by human variables such as distance to major urban centres and to highways, which were related to water pollution downstream (Barbosa et al., Reference Barbosa, Real, Olivero and Vargas2003).
Although there is no direct relationship between numbers of spraints and numbers of otters, sprainting activity reflects changes in the distribution of otters (Chanin, Reference Chanin2003) and increases with otter numbers (Lanszki et al., Reference Lanszki, Hidas, Szentes, Révay, Lehoczky and Weiss2008) and habitat use (Clavero et al., Reference Clavero, Prenda and Delibes2006). In general, 200–600 m surveys are effective for ascertaining otter presence (Chanin, Reference Chanin2003); to assess marking intensity, however, constant-length surveys are recommended (Balestrieri et al., Reference Balestrieri, Remonti and Prigioni2011). Nonetheless, the numbers of spraints and sprainting sites did not differ from those recorded in surveying a 60 km stretch of the River Drinos (Hysaj et al., Reference Hysaj, Bego, Prigioni and Balestrieri2013), and corresponded to those reported for peripheral areas of the otter's range in Italy (Prigioni et al., Reference Prigioni, Balestrieri, Remonti, Sgrosso and Priore2006; Balestrieri et al., Reference Balestrieri, Remonti, Smiroldo, Prigioni and Reggiani2008). Although two surveys are insufficient to assess any trend in the otter population in Albania, variation in marking intensity suggests that changes in socio-economic policies during 1986–2013 may have had a greater impact on otter numbers than on the large-scale distribution of the species.
During the early transitional period from communism to democracy, lack of control and effective management policies exacerbated the environmental problems inherited from the previous centralized planning political model. Privatization of land has led to the urbanization of plains and the conversion of land-use from broadleaved forests to herbaceous crops in hilly areas, with detrimental effects on the environment (Jansen et al., Reference Jansen, Carrai, Morandini, Cerutti and Spisni2006). Uncontrolled fishing, deforestation, river-bed excavation and pollution of rivers by wastewater from tanneries and petroleum refineries are still considered major threats to the otter (Cullaj et al., Reference Cullaj, Hasko, Miho, Schanz, Brandl and Bachofen2005; Hysaj et al., Reference Hysaj, Bego, Prigioni and Balestrieri2013). The burning of urban waste in the open air is the main contributor of PCDD/PCDF (polychlorinated dibenzo-p-dioxin and dibenzofuran) emissions into freshwater (Kleger et al., Reference Kleger, Kuncová, Mansaku–Meksi, Selfo and Baraj2006).
Nonetheless, since the beginning of the 21st century there has been increased awareness of environmental problems, and some progress has been made in policy and legislation. A National Strategy and Action Plan for Biodiversity was approved in 2000, and in 2007 the otter was included in the Red Data Book of Albanian fauna, categorized as Vulnerable (Cullaj et al., Reference Cullaj, Hasko, Miho, Schanz, Brandl and Bachofen2005; Hysaj et al., Reference Hysaj, Bego, Prigioni and Balestrieri2013), although it can still be hunted under permit, to prevent damage to fisheries. The Stockholm Convention was signed in 2001 and a National Implementation Plan for the reduction and disposal of persistent organic pollutants was approved in 2006.
The otter is still widespread in Albania, and effective pollution management policies may favour the reinforcement of otter populations in the future, as has occurred in several European countries (Ruiz-Olmo et al., Reference Ruiz-Olmo, Lafontaine, Prigioni, López-Martín and Santos-Reis2000). As shown by the systematic monitoring of otter populations in the UK (e.g. Strachan, Reference Strachan2007), annual otter surveys, complemented by effective statistical analysis of pollution data collected by environment protection agencies, can identify habitat changes and guide conservation programmes.
Acknowledgements
This study was conducted within the framework of the Institutional Support to the Albanian Ministry of Environment, Forest and Water Administration (MoEFWA) for Sustainable Biodiversity Conservation and Use in Protected Areas project by IUCN, funded by the Italian Ministry of Foreign Affairs, General Directorate for Cooperation Development, in collaboration with MoEFWA. We thank the project manager, Andrea Ghiurghi, for supporting and facilitating the research, and the trainees of the Animal Ecology Units of IUCN's project, specifically Tonin Macaj, Eduard Gajtani, Zeqir Ujka, Festim Broja and Elvin Shala (APA Shkoder) of the Buna River Protected Landscape, and Lavdim Qoshi, Enver Koçi, Mitat Bicaku (Librazhd Forest Office), Eriola Katjai, Luljeta Koçi and Bledar Pepa (Elbasan University), and Bledi Oxha (PPNEA NGO) of Shebenik–Jabllanice National Park. We thank Etleva Hysaj and Ferdinand Bego for help with fish identification and promoting research on otters in Albania. Lesley C. Wright (IUCN Species Survival Commission Otter Specialist Group) and Prof. Christopher F. Mason kindly revised the English.
Biographical sketches
Alessandro Balestrieri and Claudio Prigioni's research activities focus mainly on the behavioural ecology and conservation of meso-carnivores. They are also involved in the management of invasive coypus in northern Italy. Simone Messina participated in this research for his undergraduate thesis. Francesca Pella is involved in conservation research, wildlife management, and capacity building in developing countries. Nicola Saino's main research interests are sexual selection, the evolution and consequences of maternal effects, parent–offspring communication and sex allocation, and the impacts of climate change on migration phenology and population dynamics. Mauro Fasola's research interests range from demography and foraging ecology of birds, to behaviour of reptiles, environmental contamination, and phylogeography.