A NEW METHOD OF COUNTING PHOCA VITULINA SSP. STEJNEGERI (PHOCIDAE, CARNIVORA) ON THE COMMANDER ISLANDS (RUSSIA)

The Kuril Seal Phoca vitulina ssp. stejnegeri is a permanent inhabitant of the coast of the Commander Islands. The ecology of the Kuril Seal, and in particular, the features of its life cycle, the population structure and the level of fertility are poorly known. Most publications are devoted to the study of populations and spatial distribution of this animal on the islands. From June to October 2017, the current occurrence, features of spatial distribution, reproductive success of Phoca vitulina ssp. stejnegeri, seasonal dynamics of the number and the comparison of the accuracy of various methods of its counting have been carried out on Bering Island, Toporkov Island, Ariy Kamen Island, and at a site Bobroviye Kamni on Medniy Island. The material on the number and spatial distribution of the seals was collected using an unmanned aerial vehicle, quadrocopter (drone). In this study, the Kuril Seal was first counted on the islands by aerophotography. The determination of the population size has been carried out during the active reproduction phase of the Kuril Seal– from mid-June to 10th July 2017. Our study showed that the accounted abundance of the Kuril Seal was 1543 individuals, including 343 pups on Bering Island and Toporkov Island. The abundance of Phoca vitulina ssp. stejnegeri on Medniy Island has been counted using the extrapolation method. It was estimated at 689 individuals. And the total number of this animal on the Commander Islands was 3344 individuals. For the first time, we obtained data on the proportion of pups (22.2%) of the Kuril Seal on the Commander Islands in rookeries during the breeding season and the specific birth rate (17.4%) was calculated. Up to 69.3% of the total number of Kuril Seals is present in large rookeries and only a small portion is presented by small groups or individuals. The proportion of pups born in large rookeries is 72%. We found an increase of the Phoca vitulina ssp. stejnegeri abundance in the rookeries during the moult period (August). The abundance of individuals continued to grow in some rookeries in September. A comparison of the accuracy of the visual, traditional on the Commander Islands, method of animal counting with data obtained using the aerophotography method showed that in the case of counting from the boat, an observer can register an average of 67.7% of adult individuals and 24% of pups. In case of observation from the first seaside terrace of a shore, the observer can see 39.6% of individuals (adults and pups). And in case of observation from the slopes (height of 17 m and higher) the observer can see up to 62% of individuals (adults and pups). To discuss the long-term dynamics of the number of Kuril Seals on the Commander Islands it is required to conduct a detailed analysis of all available data, as they were collected by using different methods, which are not correctly comparable with the newly obtained results.


Introduction
Phoca vitulina Linnaeus, 1758 is a species distributed circumpolarly. It is presented by five subspecies: Phoca subsp. concolor De Kay, 1842, Phoca subsp. mellonae Doutt, 1942, Phoca subsp. richardii Gray, 1864, Phoca subsp. stejnegeri, and Phoca subsp. vitulina Linnaeus, 1758(Lowry, 2016. The subspecies Kuril Seal Ph. subsp. stejnegeri inhabits the Far Eastern waters of Russia, including the Commander Islands. Taking into account the great subspecies diversity and wide range -from the Atlantic Ocean basin to the Pacific Ocean -Phoca vitulina has developed in different parts of the range specific biological features, both in terms of breeding time, moulting time, use of the coastal strip, adaptation to tidal activity of the ocean, and many others. The Kuril Seal is listed in the Red Data Book of the Russian Federation. The Kuril Seal depends extremely on the daily activity rhythms on tides. So, the animals come out for a rest on bare reefs or beaches at low tide. However, the number of the seals released to rest is significantly affected by various environmental factors, such as wind direction and speed, sea waves, precipitation, temperature and even cloudiness (Pauli & Terhune, 1987;Brasseur et al., 1996;Grellier et al., 1996;Reder et al., 2003;Simpkins et al., 2003;Granquist & Hauksson, 2016). The number of seals on dry land depends on the season of the year (e.g., Granquist & Hauksson, 2016). In addition, depending on the local morphological features of the shores, Phoca vitulina can form rookeries on the reefs, which are far from the bedrock coast. This makes it extremely difficult to count the number of individuals from the shore.
The most effective method for obtaining accurate data on the number of Kuril Seal is aerophotography using light-engine aircraft and helicopters (Thomp-son & Harwood, 1990;Cronin et al., 2004). During a flight the researchers make photos and determine the abundance of animals using photographs.
The abundance investigation should be carried out at a certain time. According to Thompson et al. (2005), the best time for investigation is the period of two hours before the maximum low tide and two hours after the maximum low tide. According to Granquist & Hauksson (2016), the most optimal time is the period of two hours before the maximum low tide and three hours after the maximum low tide. In some areas, the most optimal time is a period of 1.5 hours after the maximum tide (London et al., 2012). Thus, to plan the countings it is necessary to know the peculiarities of a rookery formation on a certain place.
As not all individuals can be on the coast at the time of low tide, to determine the total number of individuals, it is necessary to use the corrective amendments for the unaccounted part of the population. These correction factors may differ in various habitats. For instance, in the study of Phoca vitulina in Norway, Bjørge et al. (2007) used the correction factors from 1.25 to 1.75 at different locations (i.e., 57% of the whole population in the rookery). In California, the correction factor varied from 1.54 to 2.86 depending on the habitat (Harvey & Goley, 2011). Tagging of seal individuals is used to determine the correction factor.
In the Russian Far Eastern waters, Phoca vitulina ssp. stejnegeri inhabits the Kurile Islands, small area of the southeastern coast of Kamchatka Peninsula and the Commander Islands. Moreover, according to general estimates, about half of its Far Eastern population lives on the Commander Islands (Burkanov, 1988;Trukhin, 2009;Zagrebelny & Fomin, 2012, 2014.
A large number of publications on Kuril Seal is devoted to the determination of the abundance of Phoca vitulina ssp. stejnegeri on the Commander Islands (Marakov, 1962(Marakov, , 1964(Marakov, , 1967Pinigin & Pryanishnikov, 1972;Mymrin et al., 1979;Burdin et al., 1991;Zagrebelny & Fomin, 2012, 2014. In addition, there are many unpublished data on the countings of the abundance of Kuril Seal (Annual report on the work, 1976-2004Chronicle of Nature, 1995-2010. In most cases, the abundance of Phoca vitulina ssp. stejnegeri was investigated from the coast and only in a few cases -from a boat. Considering the characteristics of the Kuril Seal rookeries and the coastal zone, we had certain difficulties arising in determining the accurate number of individuals. As a result, there was a lack of data on its reproductive success on Commander Islands. In addition, undoubtedly, mistakes of individual researchers influenced the counting results. The spatial distribution of Phoca vitulina ssp. stejnegeri along the coasts of the Commander Islands is a well-studied issue. Therefore, the dynamics of the coast use by Kuril Seals can be traced since 1960s till present. On the contrary, data on the general biology of the Kuril Seal, its ecology on the Commander Islands, are not enough. So, only two publications (Marakov, 1967;Burdin et al., 1991) contain brief data on the breeding terms of Phoca vitulina ssp. stejnegeri on the Commander Islands. The interannual and seasonal dynamics of the Kuril Seal abundance on the individual rookeries was considered in Burdin et al. (1991).
Apart from Phoca vitulina ssp. stejnegeri, there is an insignificant number of Phoca largha Pallas, 1811, known from the coasts of the Commander Islands. Both animals are actually indistinguishable during the field studies. Traditionally, both species are studied together on the Commander Islands. In our study, we did not make any distinction between them either.
Due to the peculiarities of the Kuril Seal biology and numerous data on its abundance, we aimed to obtain accurate data on the abundance of Phoca vitulina ssp. stejnegeri, including data on the number of pups. It is needed to obtain data on the birth rate, to determine correction factors for visual counting of animal abundance, and to calculate the absolute size of the Phoca vitulina ssp. stejnegeri population on the Commander Islands.

Material and Methods
Data were collected from 12.06.2017 to 18.10.2017. To determine the abundance, the data collected from 25.06.2017 to 10.07.2017 (breeding period of Phoca vitulina ssp. stejnegeri) were used, because they are more closely confined to the reefs and their number is at the maximum. Data were collected during the whole study period to compare the accuracy of different methods of counting and ascertaining of the seasonal (during different physiological stages of the life cycle of Phoca vitulina ssp. stejnegeri: the breeding period and the moult period) dynamics of the animal abundance in the largest areas. All investigations have been carried out by author.
To take aerophotographs of the seal rookeries, we used an unmanned aerial vehicle (quadrocopter). The «Phantom 4 Pro+» quadrocopter (drone) was used to determine the number of individuals.
The focal distance of the camera lens is 8.8 mm. During the survey, the flight altitude of the quadrocopter was 60 m, and it was chosen empirically. Thus, in case of photographing from a height of 70 m, the quality of the final photographs was not always suitable for counting the number of Kuril Seal individuals and detecting pups among reefs. Photographing from a height of less than 60 m caused trouble among individuals of Phoca vitulina ssp. stejnegeri due to the the noise of the working screws of the quadrocopter. In most cases, the quadrocopter flying at a height of 60 m did not cause any trouble, except in cases of almost total silence. Under these conditions, the Kuril Seals demonstrated anxiety during the flight. In this case, we raised the quadrocopter higher.
To study the rookeries available for land access, the quadrocopter has arrived from the side of the coastal terrace. For example, these areas are in the Severnoe and Severo-Zapadnoe rookeries, along the eastern coast of Bering Island (Bolshaya bay, Staraya Gavan bay, Buyanovskiy reef, etc.), Tonkiy cape ( Fig. 1). Other rookeries have been surveyed from the seaside. For this purpose, we launched the quadrocopter from the boat at a distance of several hundred metres of the rookeries. Toporkov Island was examined either from the boat or by landing on it from the village on the traditional place. The rookery at Skovoroda cape was surveyed from the sea and by the quadrocopter launched from Cherniy cape. During the study period, the whole coastline of Bering Island, Toporkov Island and Ariy Kamen Island has been surveyed. Phoca vitulina ssp. stejnegeri was not found on Ariy Kamen Island.
Before the launching of the quadrocopter, Phoca vitulina ssp. stejnegeri has been investigated from the coastal terrace, slope or boat to estimate the accuracy of the visual counting method. For visual counting, we used binoculars «Canon 12 × 36 IS III» and a telescope, «Baush and Lomb 15-60 × 60» with a tripod. The counting of Phoca vitulina ssp. stejnegeri by the telescope has been carried out with magnification power at 20 times.
All countings of Phoca vitulina ssp. stejnegeri were made in periods with the maximum level of low tide, provided the appropriate weather conditions: wind speed -up to 10 m/s, no precipitation (fog, rain). In autumn, the time of the day was a restriction of the quadrocopter use in days with maximum low tides. The greatest low tides are shifted to the early morning hours in autumn. Hence, twilight did not allow us to use the quadrocopter for counting Kuril Seals.
All investigations have been carried out within a 4-hour period: two hours before the low tide and two hours after, in accordance with the behavioural features of Phoca vitulina ssp. stejnegeri on the rookeries (Thompson et al., 2005;Granquist & Hauksson, 2016). However, due to a fairly long coastline, it was not always possible to be near the rookery at the necessary time. Nevertheless, 82.5% of the total number of surveys were performed within the established 4-hour window.
In total, 52 flights with a duration of 874 min. and a length of 203.4 km were performed on Bering Island, three flights with a duration of 59 minutes and a length of 14.5 km -on Medniy Island, seven flights with a duration of 149 min. and a length of 24.3 km -on Toporkov Island. A total of 6725 photographs were made.  10.07.2017. The total number of Phoca vitulina ssp. stejnegeri individuals was 1200 adults and 343 pups (Table 1 and Table 2).

The number of Phoca vitulina ssp. stejnegeri on Bering Island and Toporkov
The proportion of pups on the rookeries was similar for both islands, with an average of 22.2%. The number of Kuril Seal individuals on Toporkov Island is 19.6% of the number on Bering Island.
Obtaining data on the number of any species is aimed to estimate the current state of a population for a long-term aspect. For this purpose, we attempted to analyse numerous data on the abundance of Phoca vitulina ssp. stejnegeri on the Commander Islands. Vaksel (1940) mentioned the first record of Kuril Seals on the Commander Islands. The author only briefly reported that «there is a great number of Kuril Seals» on the islands during the wintering of the Vitus Bering expedition in 1741/1742. Grebnitsky (1882) also pointed out that earlier there was a large number of the animals, but they were almost exterminated. The first quantitative data on the Phoca vitulina ssp. stejnegeri abundance can be found in Barabash-Nikiforov (1935). The author noted that on the site Bobroviye Kamni on Medniy Island, there were up to 100 or more individuals together, and only one rookery remained of the large number earlier. On the contrary, Sergeev (1938) reported that Kuril Seal formed several permanent rookeries on each of the islands. Ilyina (1950) made only the general estimation and noted that earlier from the Kuril Seals had been recorded a much larger number and that a part of their rookeries was emptied. Ilyina (1950) did not give data on the number of Phoca vitulina ssp. stejnegeri individuals, but provided information on the seal hunting. Marakov (1962) provided the first data on the total number of Phoca vitulina ssp. stejnegeri individuals on the Commander Islands. The author estimated the number of the Kuril Seals at 1500 individuals on the islands in the early 1960s. It was suggested that a «noticeable increase» in the number of Phoca vitulina ssp. stejnegeri individuals on the islands (especially on Medniy Island) and the new rookeries formation took place from the 1950s to the 1960s (Marakov, 1964(Marakov, , 1967. The total number of Phoca vitulina ssp. stejnegeri individuals was estimated by Marakov (1967) at 600 individuals on Bering Island. Marakov (1970) estimated the number of Kuril Seal individuals at 2000 individuals on the islands in the late 1960s. In the second half of the 1960s, Pinigin & Pryanishnikov (1972) determined its number at 1394 individuals on Bering Island, and 703 individuals on Medniy Island.
In the 1970s the counting of Phoca vitulina ssp. stejnegeri on the Commander Islands began to be carried out more regularly by the staff of the Commander watch control station of the Kamchatka Basin Directorate for Fisheries, Conservation and Reproduction of Fish Resources. So, according to their investigations, the number of the Kuril Seals was 1414 individuals on the islands in 1973 (700on Bering Island and 714 -on Medniy Island). At the same time, Marakov (1974) indicated that there were 2000 seals on the islands (1300 on Bering Island and 700 on Medniy Island), while Marakov (1978) noted the number of Phoca vitulina ssp. stejnegeri as 2500-2700 individuals, although it should be noted that the counting of Kuril Seals had not been carried out by exactly this author. In 1975 on the Commander Islands, according to Mymrin et al. (1979) the number of Phoca vitulina ssp. stejnegeri was 1612 individuals, while according to Goltsev & Bukhtiyarov (1978) there were 2300 individuals. The largest number of the Kuril Seals on the islands was recorded in 1978 -2816 individuals (Annual report on the work, 1976-2004). In general, in the 1970s the estimated number of Kuril Seals varied from 1402 individuals (in 1977) to 2816 individuals (in 1978).
In the 1980s there was not any complete counting of the number of Kuril Seal individuals on the Commander Islands. Numerous countings have been carried out in some large areas, most often in the Severnoe and Severo-Zapadnoe rookeries. The total number of Phoca vitulina ssp. stejnegeri individuals was determined on the basis of the average long-term data and the results of surveys on the rookeries surveyed regularly (Annual report on the work, 1976-2004; Burdin et al., 1991). Thus, in this period there were only estimated data on the abundance of Kuril Seals. So, during this period the total number of Phoca vitulina ssp. stejnegeri was estimated at 2000-2500 individuals, including Bering Island (1500-2000 individuals) and Medniy Island (500-1000 individuals).
In the 1990s there was not any complete counting of the number of Kuril Seal individuals on the Commander Islands either, and data on its abundance were evaluative. Only in 1998 there was a complete survey of Medniy Island, where the total number of Phoca vitulina ssp. stejnegeri was 1224 individuals (Chronicle of Nature, 1995Nature, -2010Zagrebelny & Fomin, 2012). In general, in the 1990s the total number of Kuril Seals was estimated at 2200-3024 individuals on the Commander Islands, including Bering Island (1600-1800 individuals) and Medniy Island (600-1224 individuals).  belny & Fomin, 2012belny & Fomin, , 2014. In general, during this period estimates of the Phoca vitulina ssp. stejnegeri abundance were 3000-4128 individuals, including Bering Island (1900-2827 individuals) and Medniy Island (1000-1555 individuals) (Vyatkin & Burkanov, 2006;Burdin et al., 2009;Zagrebelny & Fomin, 2012, 2014Chronicle of Nature, 1995-2010. In the 2010s, the countings of the Phoca vitulina ssp. stejnegeri abundance on the Commander Islands were conducted several times. So, in 2012, a boat counting was taken only around Bering Island (Chronicle of Nature, 1995Nature, -2010Zagrebelny & Fomin, 2014) where the number of Kuril Seals was 2031 individuals. In 2015, a boat counting has been carried out together with investigations of Enhydra lutris Linnaeus, 1758. It resulted in estimating the number of Phoca vitulina ssp. stejnegeri at 1079 individuals on Bering Island (including Toporkov Island and Ariy Kamen Island), and 396 individuals -on Medniy Island (Differentiated Chronicle of Nature, 2016).
As countings has not been conducted on Medniy Island in 2017, we tried to determine the total number of the Kuril Seal on the islands using extrapolation method. According to the results of previous studies (Mymrin et al., 1979;Goltsev & Bukhtiyarov, 1978;Burdin et al., 1991;Burkanov, 1988;Burdin et al., 2009 To calculate the total number of the seals, including those outside the rookery during low tide and thus not recorded, we used the formula N = 2n/(0.521 + 0.698) suggested by Thompson et al. (1997). As a result, we get the total (absolute) number of Kuril Seals on the Commander Islands -3344 individuals. As an assumption, we suggested that all newborn pups are on the rookeries and, accordingly, the calculation was made only on the basis of the number of adults plus the number of pups.

Spatial distribution of the rookeries of Phoca vitulina ssp. stejnegeri.
During the study period the entire coast of Bering Island and also Toporkov Island and Ariy Kamen Island have been investigated. In addition, the site Bobroviye Kamni on Medniy Island was examined twice. The collected material allowed us to estimate the spatial distribution of the rookeries of Phoca vitulina ssp. stejnegeri on the islands and to reveal the places, where the Kuril Seal forms the largest populations. One of the largest rookeries of Phoca vitulina ssp. stejnegeri is located on Toporkov Island during the breeding season (Fig. 1).
On the five largest rookeries, 69.2% of the total number of adult Phoca vitulina ssp. stejnegeri individuals is gathered on Bering Island and Toporkov Island (Table 3).
On Medniy Island the rookery was surveyed only at the northernmost tip of the island -site Bo- In addition to large rookeries of Phoca vitulina ssp. stejnegeri on Bering Island with populations containing more than 100 individuals, there were several sites with populations of several dozens of individuals. Among these are rookeries on Severo-Zapadniy cape, Kazarmeniy cape, Ariy Nepropusk cape and Tolstiy cape.
Apart from the collective rookeries along the seacoast, there are also single seals resting on reefs. So, most records of single animals were registered along the eastern coast of Bering Island from Starogavanskiy cape to Uyum bay. In total, there were 20 single animals. They form only 1.7% of the total number of Phoca vitulina ssp. stejnegeri individuals. Thus, the main part of the Phoca vitulina ssp. stejnegeri population on Bering Island and Toporkov Island is gathered in large groups (Fig. 3).  In addition to the counting of the animals aged 1+ years, the pups were recorded. Data on the spatial distribution of pups are shown in Fig. 1. Similarly to adult animals, the majority of pups were in five large groups (Table 3).
Forty-two pups were registered at the site Bobroviye Kamni on Medniy Island. The total distribution of Kuril Seal pups is shown in Fig. 3 ranked by number of individuals. Thus, as can be seen in Fig. 3, the vast majority of pups of 72% was found in the large rookeries.
In the largest clusters, the birth rate of the Kuril Seal varied from 18.7% (Buyanovskiy reef) to 28.2% (reefs between Staraya Slastnaya bay and Startcevskaya Shaiba bay), with an average of 23.6%. In general, the birth rate was 22.2% for all the rookeries of Bering Island and Toporkov Island.

Seasonal dynamics of the Kuril Seal population in the rookeries.
During the summer -autumn period, Toporkov Island, reefs between Tonkiy east and Prolivnoy cape, Starogavanskiy cape, Buyanovskiy reef and Bobroviye Kamni were investigated repeatedly to determine the seasonal dynamics of the Kuril Seal population on the rookeries. The results of the survey are presented in Table 4.
There was a significant increase in the number of the seals in three rookeries, surveyed in the first two decades of August (Table 4). Later, a decline in the number of individuals was noted in three rookeries, while on Buyanovskiy reef the number of individuals continued to increase until the first decade of September. The number of Kuril Seal individuals has increased almost twice on site Bobroviye Kamni.

Comparison of the accuracy of various methods of Phoca vitulina ssp. stejnegeri counting.
The accuracy of visual counting from a boat and from the coast was compared with the data obtained using the quadrocopter. The counting of Phoca vitulina ssp. stejnegeri has been carried out using the RIB boat, where the level of the eye height of an observer is approximately at 2.0-2.5 m above water level. The distance, from which the countings were carried out, differed slightly near the various reefs due to the peculiarities of the reef slabs location, depths near them, the location of the algal fields that did not allow them to approach closer, the behaviour of Phoca vitulina ssp. stejnegeri due to the wind direction. Data on results of the countings and observation distance are presented in Table 5.
The average accuracy of the visual counting of the seals is 67.7% (Table 5). At the same time, there is a significant variability in this accuracy from 29.7% to 76% and even rediscount on 19.4% (for example, on Tolstiy cape). The counting accuracy is highly depending on the relief and morphology of the reef itself, the characteristics of the animal location on a certain day, and the light conditions.
Comparative results of Kuril Seal pups counting from a boat are given in Table 6. It can be seen that during the visual counting on average 24% of pups are accounted from the boat with variations from 0% to 66.7%. At the same time, there is not a regularity in the counting accuracy, depending on the counting distance.  A visual coastal counting has been carried out both from a height of 2-3 m (at the first coastal terrace) and from slopes at a height of at least 17 m. Table 7 presents data on the accuracy of the visual counting of Kuril Seals from a height of 2-3 m without dividing the animals into age groups.
As can be seen from Table 7, an average of 39.6% of Kuril Seal individuals can be counted on the rookeries using binoculars from the coast. A comparable accuracy level was obtained by using a telescope. Table 8 presents data on the accuracy of the visual counting of Phoca vitulina ssp. stejnegeri individuals, being obtained from a height of at least 17 m. The accuracy of the visual counting increases in case of observation from slopes. And, on average, it is 62% in case of observation using binoculars. A similar accuracy level was obtained by observation using a telescope.

Discussion
The abundance of Phoca vitulina ssp. stejnegeri. Despite of the obtained data on dynamics in Fig. 1, we should note that the curve of the number of Phoca vitulina ssp. stejnegeri does not reflect adequately the right situation with the dynamics of the Kuril Seal number on the Commander Islands due to some reasons: -Sometimes data on the number of Phoca vitulina ssp. stejnegeri individuals on the islands have not been determined as a result of direct countings. Usually the number of individuals has been estimated on the basis of certain generalisation using the maximum values for separate rookeries.
-To count the number of Kuril Seals, data obtained in different seasons of the year and, accordingly, various annual life cycles of the animal were used.   In addition, we should take into account significant fluctuations of the number of individuals in some years. For example, in 1967, the number of Phoca vitulina ssp. stejnegeri was estimated at 600 individuals on Bering Island (Marakov, 1967), while in the next year it was 1394 individuals at the same location (Pinigin & Pryanishnikov, 1972 , 1995-2010Zagrebelny & Fomin, 2012, 2014, i.e. the growth was 41.4%. It is obvious, that such an abundance growth of the resident animal is not possible. This is confirmed by results of other authors for closely related animals. For example, the annual increase of the seal number in the Phoca vitulina population in Wash (England) after a severe post-epidemic reduction was only 6%, in the population of the seal in the southeastern part of the North Sea it was 8.8-12.7%, and in the population in Kattegat-Skagerrak it was 12% (Thompson et al., 2005). In most parts of the range, the growth of the Phoca vitulina number is 12-13% per year (Harkonen et al., 2002). At the same time, the death rate of pups is high enough at the end of breastfeeding. For example, in Canada it is 12-21% (Hammill et al., 2010), while the mortality value for pups before six months is up to 43%. Estimates of the number of Kuril Seals differed even within one year enormously. For example, in 1975 the difference in the number was 688 individuals according to the data of different researchers (Goltsev & Bukhtiyarov, 1978vs. Mymrin et al., 1979. In our opinion, high fluctuations of the number of Phoca vitulina ssp. stejnegeri individuals can be explained by the methodological imperfection of the conduced surveys, but not the real values of dynamics in animal abundance. In addition, it should be noted that, probably, although not obviously, previous researchers, by estimating the total number of Phoca vitulina ssp. stejnegeri on the Commander Islands, had in mind the exactly total number of Kuril Seals, but not only the accounted one. This assumption is based on the fact that during the breeding season, only a part of the seal population is located on a rookery. And according to some studies, from 52.1% of males to 69.8% of females are located on the rookery during the breeding season (Thompson et al., 1997). And during moulting, the total number of seals will be higher with an average of 37% (Sharples et al., 2009). Consequently, the total (absolute) number of the seals will be higher than the accounted number.
Based on all the aspects discussed above, related to the available data on the animal abundance, to obtain an adequate picture on the dynamics of the Kuril Seal number, an additional generalisation and analysing of the available materials are required. Before this, there is no way to discuss an objective picture of dynamics of the Kuril Seal number on the Commander Islands.

The features of the reproduction of Phoca vitulina ssp. stejnegeri on the Commander Islands.
In addition to data on the number of Phoca vitulina ssp. stejnegeri individuals on the Commander Islands, information on breeding has also been taken into account during previous decades. However, these data are extremely limited. In particular, Marakov (1964Marakov ( , 1967 noted that the pup birth occurred until late June. Vyatkin & Burkanov (2006) gave information that Phoca vitulina ssp. stejnegeri bred in late May -June. Burdin et al. (1991) pointed out that on the Commander Islands, the breeding peak of the Kuril Seal was in the first decade of June. There are not published data on the number of Phoca vitulina ssp. stejnegeri pups on the Commander Islands.
As a result of the study, we first obtained data on the number of Kuril Seal pups for Bering Island and Toporkov Island. We attempted to count the absolute and specific birth rate of pups on the Commander Islands. Thus, we made two assumptions: 1) all newborn pups are on the rookeries, i.e. the accounted number of pups is close to the value of their absolute number in a location; 2) the ratio of the number of pups on Bering Island and Toporkov Island to the number of pups on Medniy Island is 2.24, i.e. the value that we used to calculate the total number of Kuril Seals. Then we have 343 pups on Bering Island and Toporkov Island and 153 pups on Medniy Island. As a result, the absolute birth rate will be 496 pups for the Commander Islands. Using the accounted absolute number of animals beyond one-year-old (2848 individuals), we get a specific birth rate -17.4%. According to some data (Reijnders, 1982), a normal birth rate of Phoca vitulina populations is 20-30%.
There are two rookeries (Toporkov Island and Starogavanskiy cape on the east coast of Bering Island), where massive bedding and breeding of Kuril Seals are observed. These locations were examined twice during the reproductive period. During the survey of Toporkov Island, 25 pups were found on 12 th June, and 54 pups were found on 6 th July. Thus, 29 pups were born during 24 days here. This is 53.7% of the total number of pups born. On Starogavanskiy cape there were 28 pups on 18 th June, and 48 pups on 2 nd July. Thus, the number of pups increased by 45.8%. According to these data, mass breeding of Phoca vitulina ssp. stejnegeri on the Commander Islands occurs in the second half of June, but this issue needs to be clarified.
Probably, individual pups can be born also after the end of the mass breeding. For example, a newborn pup of Phoca vitulina ssp. stejnegeri was recorded on Buyanovskiy reef on 20 th August. And another survey of the rookery (7 th September) found out one pup kept contact with the mother. Before our study, the latest case of the birth of the Kuril Seal pup was noted on Toporkov Island on 17 th July, 1990(Annual report on the work, 1976-2004. Thus, the breeding period of Phoca vitulina ssp. stejnegeri on the Commander Islands is fairly extended -from May to August. But the main part of pups are born probably in the second half of June.
Spatial distribution. The collected material makes it possible to estimate the spatial distribution of the Kuril Seal rookeries. The main large clusters were found on Toporkov Island, on Tonkiy east-Prolivnoy cape, on Starogavanskiy cape, on Buyanovskiy reef, in Startsevskaya Shaiba bay.
In the 1960s on the coast of Bering Island the largest cluster of Phoca vitulina ssp. stejnegeri was recorded in Kazarma bay with up to 200 individuals (Marakov, 1967). In comparison with the results of studies in 2005 (Zagrebelny & Fomin, 2012, 2014, in 2017 there was no large group of animals in the Severo-Zapadnoe rookery. But the number of the seals increased on Starogavanskiy cape. However, it should be noted that the counting has been carried out in late July 2005, i.e. at the beginning of the moulting, when the Kuril Seal number on the rookeries is believed to increase. In addition, the number of Phoca vitulina ssp. stejnegeri on some rookeries can significantly fluctuate within a month and during several days (Nikulin & Nikulin, 2012;our data). In this connection, it is difficult to draw unequivocal conclusions on the changes in the spatial distribution of Phoca vitulina ssp. stejnegeri groups along the rookeries. Nevertheless, the general features of the arrangement of the most significant sites of the Phoca vitulina ssp. stejnegeri concentration remain. On Medniy Island, the largest (from 100 to 300 individuals) rookery of Phoca vitulina ssp. stejnegeri was found on Bobroviye Kamni site like in all previous years since the 1930s. (Barabash-Nikiforov, 1935;Marakov, 1967). Thus, in general, there are some remaining peculiarities of the spatial distribution of Kuril Seals, but there are also differences. This issue requires a more detailed elaboration in the future.
Seasonal dynamics of the Kuril Seal number. As a result of the work carried out, seasonal changes (from June to October) in the Phoca vitulina ssp. stejnegeri number on some rookeries have been revealed. The growth of the seal number during the seasonal moult, which occurs in August, was recorded on all the rookeries that could be inspected in the first half of August. And the number of Kuril Seals continued to grow on the distinct rookeries in September. It is difficult to draw conclusions at the expense why the number increases. It is believed that the increase in the number is a consequence of the fact that a larger number of the seals are present at the rookeries. But it is also possible there is some redistribution of the seals between the rookeries. Thus, the decrease in the Phoca vitulina ssp. stejnegeri number on Starogavanskiy cape in September and the increase of its number on Buyanovskiy reef can be attributed to the spatial redistribution of animals along the eastern coast of Bering Island. However, additional research is required for an unambiguous approval.
Comparison of the counting methods accuracy. Despite the fact that Phoca vitulina ssp. stejnegeri is a permanent inhabitant of the coastal waters on the Commander Islands, detailed studies of its ecology have not been conducted. Even the estimate of the total number was carried out for the first time only in the 1950s. In general, two counting methods of the Kuril Seal number have been used on the Commander Islands -from the shore (coastal counting) and from a boat (boat counting).
Unfortunately, there is no description of the methods used by Marakov (1962Marakov ( , 1967Marakov ( , 1974 for counting the animal number. So, it is not clear whether the author counted Phoca vitulina ssp. stejnegeri from the shore or from the boat. We do not know either how often such countings were made, in which seasons, which rookeries and how many times during the year. Taking into account the round values of the seal number in Marakov's (1962Marakov's ( , 1967Marakov's ( , 1974 data, it can be assumed that the author did not give any accounted data. But the author provided an estimated intuitive number of Phoca vitulina ssp. stejnegeri on the islands.
The first data of the accounted number of the Kuril Seals are given by Pinigin & Pryanishnikov (1972). However, the authors counted the seals on Bering Island in February, and on Medniy Island in June, 1968. Thus, the data obtained in different seasons cannot be compared with the results of the surveys performed during the breeding season of Phoca vitulina ssp. stejnegeri. In addition, it is not clear how many rookery surveys were performed on Bering Island in February and what data on the number of animals were used for the publicationthe maximum for several records or average values. There is no information on the counting method of Phoca vitulina ssp. stejnegeri on Medniy Island. Further estimates of the Kuril Seal number on the islands were made on the basis of the average longterm data on the individual rookeries (Burdin et al., 1991). Thus, for the entire period of the counting, we do not have accurate data on the Phoca vitulina ssp. stejnegeri number on the Commander Islands. The exception is data (sometimes long time series) obtained on distinct places of the rookeries. But in publications or reports there is no detailed description of the methods of counting. The only description of the method of boat counting can be found in Zagrebelny & Fomin (2012, 2014. To clarify the methodological features of coastal countings, we requested N.I. Mymrin and V.S. Nikulin, who worked in the Commander watch control station of the Kamchatka Basin Directorate for Fisheries, Conservation and Reproduction of Fish Resources in 1970-1990s. And they were engaged in counting of Phoca vitulina ssp. stejnegeri. So, according to the oral report of N.I. Mymrin, who has carried out countings in the 1970s, the study of Phoca vitulina ssp. stejnegeri on the rookeries along the eastern coast of Bering Island was led from the first coastal terrace with the use of 7-fold binoculars. In 1980-1990s the counting of Kuril Seals has been carried out from the coastal line using 8-10-fold binoculars (V.S. Nikulin, personal communication). To carry out countings in the Severnoe and Severo-Zapadnoe rookeries, a telescope 30-60 was used. Extrapolations were not used. According to the oral report of E.S. Baldin, who conducted countings in the 2000s, the studies along the eastern coast of Bering Island were also performed from the first coastal terrace or from the road. Occasionally, the inspection of Starogavanskiy cape was made from a 10 m coastal cliff. In this survey, 8-10-fold binoculars were used. Only the directly counted seal number was recorded in the diary of observations, i.e. without extrapolation to the invisible hidden part.
One of our tasks was to establish the accuracy of the visual counting in comparison with the counting with photographs taken from the quadro-copter. As already mentioned, the most accurate data on the number of Phoca vitulina on the rookeries can be obtained only during aerophotography (Thompson & Harwood, 1990;Cronin et al., 2004). So, we found out that during the boat counting the observer can record on average only 67.7% of adult animals and 24% of pups. During coastal countings, on average, the observer can register only 39.6% of Kuril Seals of all age groups, if the observation is taken from the first coastal terrace, similarly to studies of previous years. In other parts of the Phoca vitulina range, the accuracy of the visual counting in comparison with aerocounting averaged 73% and depended on the number of the seals: the accuracy decreased with increasing of the number of seals (Thompson & Harwood, 1990).
However, even if we know the level of undercounting by using the traditional visual counting of Kuril Seal, it is difficult to apply the correction factor to the data obtained in this way. This is connected with the individual error of the accountant: so, the accountant can both underestimate the seals and overestimate their number. In addition, the error of the visual counting for the same accountant varies depending on the pecularities of arrangement of the seals on a rookery at different times. For example, the animals may be arranged more open and dispersed, or may also move over the protruding ridges and stones on the reefs and lie densely. Also, the lighting conditions, distance and other factors also impact on the counting accuracy. For example, in Zagrebelny & Fomin (2012, 2014 the authors noted that the counting was made at a considerable distance from the rookeries and in conditions of limited time, i.e. the rookeries were inspected for a short time. In such conditions it is extremely difficult to determine the correction factor for obtaining updated data on the number of Kuril Seals. Nevertheless, the obtained data allow us, to some extent, to correct the values of the countings performed in previous years by using the visual method.

Conclusions
For the first time, Phoca vitulina ssp. stejnegeri was counted using aerophotography on the Commander Islands. This method gives the most objective data on the number of animals. Data obtained using an unmanned aerial vehicle on the current number of Kuril Seals do not allow us to estimate the long-term dynamics of its number on the islands. This is due to the fact that since the 1950s, all studies of the Kuril Seal number have been performed by using the visual method. And the general estimates of its number on the islands were most often given rather on the basis of some expert estimates, than calculations. In order to include the whole data on the Phoca vitulina ssp. stejnegeri number on the Commander Islands in assessment of the historical dynamics of its abundance, it is first necessary to conduct a thorough analysis of the counting data. Such analysis should include both an estimate of the Kuril Seal number on each particular rookery, and its dynamics during the season and between years.
To determine the absolute number of Phoca vitulina ssp. stejnegeri on the Commander Islands, we made assumptions about the level of the Kuril Seal number on the rookeries at low tide, based on available literature from other parts of its range. However, the local populations of the animal may have its own specific features. To clarify this issue, it is required to tag the animals. Labelling combined with using satellite transmitters will also make it possible to assess the nature of the water area use on the islands by Phoca vitulina ssp. stejnegeri, and the degree of their confinement to specific rookeries, and many other aspects.
Also, for the first time, data on the number of pups on the Commander Islands were obtained. The reproduction level was close to that in other parts of the Phoca vitulina range. Simultaneously, the continuation of such studies is required in order to exclude possible interannual fluctuations and counting errors.
The study showed that Phoca vitulina ssp. stejnegeri uses traditionally the coasts of the Commander Islands. So, the largest rookeries are located in the same places where they were known during the previous decades. Such a traditional arrangement of the coast by the animal can be used for rapid assessment of its number on the rookeries, by surveying only the largest rookeries, taking into account what proportion they occupy of the total number of Kuril Seals.
The seasonal dynamics of the Kuril Seal number has not been adequately studied. Nevertheless, it has been shown that during the seasonal moult, the number of the seals increases on the rookeries. Although, this does not happen everywhere. Therefore, research on this topic should be continued in future.
An important result of our study was the assessment of the accuracy of the visual counting method, traditionally used on the Commander Islands. A correction factor was obtained for the Kuril Seal counting. Taking into account certain limitations, it can be used to adjust or discuss the available historical data.
In general, taking into account the biological pecularities of the Kuril Seal, it can be concluded that Phoca vitulina ssp. stejnegeri is a complex object for research and estimation of its abundance. Undoubtedly, one season studies are not enough to draw serious conclusions about the state of its population.