Harvest history

In Russia, marine mammal hunting was a major component of traditional subsistence activities and was common in many coastal indigenous peoples’ communities including those located in the Russian Far East along the Sea of Okhotsk. A number of ethnic groups relied on marine mammals as a major source of food and essential materials. For instance, in the beginning of the 20th century, 63% of Koryak households on the Kamchatka peninsula were engaged in seal or whale hunting [20]. Nivkhi on Sakhalin and in the Amur River estuary were known to hunt sea mammals, especially seals and sea lions, using floating harpoons and clubs from boats or from the shore. Negidal, Evens, Nivkhi and some other indigenous peoples have hunted marine mammals off the coast of the Sea of Okhotsk since the 18th century using bidarrahs (a multi-seat baidarka), kayaks and other types of boats [21].

During an intensive period of growth of settlements in the early to mid-20th century, commercial sealing started as coastal seal hunting and then became well developed with construction of specialized sealing vessels in the late 1950s and early 1960s [7]. The commercial seal harvest established in the Russian Far East in 1932 was unregulated until 1968 for all the species hunted. Until the late 1960s, harvests were mainly for blubber and skin and took place predominantly on molting grounds [22]. Due to both a lack of harvest limits and non-selectivity of catches, harvest data were extensively used to draw conclusions on a populations’ status, as catches were presumed to reflect seal abundances [9]. Scarce information available on seal catches before the 1950s suggests that aboriginal seal hunting did not exceed 25,000–35,000 individuals for all species combined per year in the northern part of the Sea of Okhotsk [6]. According to Fedoseev’s estimates [9], in the years of intensive unregulated sealing, 1955 to 1968, the average annual catch was about 78,500 ringed seals, 13,000 ribbon seals, 10,100 bearded seals, and 4,600 spotted seals.

By 1963, high harvest rates led to depletion of seal stocks and the catches started declining [6]. In 1968, hunting limits were set on vessel- and shore-based harvests, wherein hunting of ringed and spotted seals was permitted only from May to July [23]. Ribbon seal harvests were restricted to 7,000 per year due to stock decline. At the same time, Fedoseev [22] pointed out that the harvest limits introduced initially were biased as they did not take into account unbalanced exploitation of the regional structure in seal populations; dense reproductive concentrations of seals were under major pressure and continued rapidly toward depletion.

We interpret a trough in total harvest rates appearing in the second half of the 1970s (Fig 2) to be a consequence of a science-based decision made by state authorities to reduce total allowable catch rates for all seal species in the Sea of Okhotsk. The purpose of this was to allow the populations to recover from a period of uncontrolled exploitation followed by several years of high harvest quotas. The quotas were substantially increased again in 1982.

In 1968–1975 the Soviet government issued a series of decrees ordering the development of fur-oriented sealing. This caused harvests to become highly unsustainable: small schooners were unable to keep the meat on board so they discarded it, and because the main target of sealers shifted towards molted individuals, a major portion of seals killed were thrown overboard due to insufficient fur quality. Hunting of bearded seals was virtually stopped due to the low value of its pelt [5, 6, 22].

In the early 1990s, after the collapse of the Soviet Union, the state ceased subsidizing the sealing industry and the seal harvest collapsed. Since 1995, sealing has been conducted in the Sea of Okhotsk only by enterprises representing the interests of indigenous peoples and their total harvest did not exceed 1–4% [7] of the quota of approximately 50,000 seals for all species combined, allocated to the Magadan Region in 1995–2005. Starting in 2006, Total allowable seal catches were reduced drastically to 7,800 in 2008 and 2,100–2,200 in 2009–2012. After 2012, ice-associated seal species were moved from "Total allowable catches" system for the Sea of Okhotsk to "Possible catches" which is a less strict system used for less commercially valuable species with low harvest interest [24–44].

The shore-based harvest pressure on seals has always been lower compared with vessel-based harvest pressure. We found records on shore harvest rates in Okhotskrybvod reports from the study period (see Table 2). Also, Trukhin [1] summarized those rates for a period from 1947 to 2004. For the years in which the data on both types of harvest were available, the mean contribution of shore harvest to the overall seal take in the Sea of Okhotsk was 0.26 (range 0.035–0.893) for ringed seals, 0.28 (range 0.093–0.721) for bearded seals, and 0.11 (range 0.004–0.485) for spotted seals. Ribbon seals were not harvested from shore (except low numbers reported in catches in 1989–1994 and 2004). To the best of our knowledge during the 20th century, 353,003 ringed, 29,329 bearded, and 24,697 spotted seals were harvested from shore [1].

Reliability of log book data

We consider log book data analyzed to be a representative sample of the total harvest effort for the study period: based on official records on total number of vessels working in 1972–1994 provided by the Russian Federal Research Institute of Fisheries and Oceanography (VNIRO) and reported by Okhotskrybvod, we possess data from approximately 94% of the sealing cruises. Nevertheless, we encountered certain discrepancies between catch rates for particular species recorded as the harvest progressed, and harvest rates reported later in the literature [1–7]: since the sources of information were often not reported in the scientific publications, we believe that the values of harvest rates might have been partially estimated based on incomplete information available to the authors at that time. While analyzing raw log book records, we also noted that certain numbers of seals initially recorded as taken for their pelts were later moved to the record field reflecting that they were taken for their meat, or vice versa. Some low numbers of mathematical or typing errors were found in the logs. Mapped seal catch locations (Figs A-P in S1 File) based on log book data sometimes appeared to follow straight lines going from north to south or from east to west. This is an artifact of the specifics of the geographical coordinate format provided in the sealing log books, where latitude and longitude were rounded to two decimal places due to the absence of more precise geolocation systems.

Harvest and fishing vessels working at sea were obliged to accommodate an enforcement officer on board if one was provided. Based on log book records, officers were only present on some cruises and could work only on one of the skiffs, whereas most vessels had several skiffs. This officer, generally either a state official or biologist from a research institution, was empowered to enforce seal harvest regulations according to sealing areas and quotas as well as proper reporting. State officials could be involved in harvests directly as hunters (and therefore receive additional income). In these cases, we assumed that the reliability of the inspection results could have been compromised due to extra work duties performed and potential conflicts of interest.

We did not discern any relationship between the presence of an enforcement officer on a vessel and the accuracy of the records, or the consistency in reporting of harvest losses. For instance, even though vessels had an officer on board, daily harvest rates for all species were apparently rounded (100, 150, 200, etc.) on some cruises during 1976. We also found several instances when the same person was listed as an enforcement officer on two different vessels simultaneously working at sea. Although we believe that the discrepancies described above do not diminish the overall value of the data preserved in the log books, they should be borne in mind when interpreting our results and drawing specific conclusions.

Temporal variation in catches

We hypothesized that harvest start dates differed somewhat depending on ice coverage in April. Although we expected harvest to start later in years with heavier ice conditions, the model outcome showed the opposite trend. According to the recollections of the authors of the present study who participated in the harvest cruises, use of HFVs removed the limitations associated with ice conditions and the vessels began the harvest depending on organizational and technical aspects, mainly vessel maintenance, expedition supply, and crew recruiting. At the same time, the lack of a significant effect of ice conditions on catch rates implies that within observed ranges of ice coverage in 1972–1994 sealers always managed to take their quotas, apparently due to abundant seal populations.

In the Sea of Okhotsk we did not find any pronounced trends in the frequency of high ice years or low ice years, neither during the study period nor in the following years. Based on our classification, the proportion of high ice years slightly decreased from 0.14 to 0.09, in low ice years from 0.41 to 0.32, and in medium ice years increased from 0.45 to 0.59 when comparing data from 1972–1994 with data from 1995–2013. Finer scale data are needed to demonstrate a possible effect of ice conditions on seal population abundances and harvest success.

The beginning of sealing season in 1972–1994 typically coincided with peak pupping season for ringed seals, and early pupping season for ribbon, bearded and spotted seals in the Sea of Okhotsk [45–48]. The presence of white-coated pups reported in catches in April to mid-May for ringed, spotted and ribbon seals corresponds with the general timing of molt in pups of these species [45–48]. Bearded seal pups in their natal coats, which is brown, are unlikely to be found in the catches in high numbers as many of the pups complete at least part of their molt in utero [49]. Single records of natal-coated bearded seal pups in catches from June-July might be a result of delayed parturition of individual females (possibly young) or data entry errors.

Ringed seal CPUE increased gradually from the beginning of harvest to reach its maximum level only in mid-June, later than the CPUE of the other three seal species (see Fig 3A). Ringed seals occur primarily in shorefast ice regions not easily accessible to sealing vessels in April; this explains the larger CPUEs later in the season when ice started melting. Low catch values early in the 1972 and 1974 seasons may be due to the use of smaller wooden boats incapable of penetrating the compact ice zones. Another reason for low catches might be that at least some pups are born in under-snow lairs (especially for ringed seals breeding on shore-fast as opposed to the drifting ice in the Sea of Okhotsk; [9, 50, 51]). Those seals would become available for sealers only after the snow melts. After reaching its maximum rate, ringed seal CPUE remained more or less stable until the end of the harvest season.

Ribbon seal CPUE (see Fig 3B) followed a more or less symmetric curve, peaking in late May when these seals transition from nursing to molting season. Ribbon seals spend comparably more time on floating ice in the ice-edge zone [45] than other ice-associated species. They were a conspicuous and important harvest species throughout the season. The gradual decline in observed catches in July could be explained by a highly-dispersed distribution that did not provide harvest concentrations of sufficient size.

Low bearded seal CPUE, in comparison to the other species, could be related to the ability of bearded seal females and pups to escape into the water during the nursing season, and a generally lower interest in this particular species from hunters. Bearded seal catches were slowly increasing, presumably due to the beginning of the molt, from late May to early June and continued to do so until the end of the harvest season (see Fig 3C).

Spotted seal harvest seemed to have been more pup-oriented compared with other species as it primarily took place on pupping grounds. Outside the breeding season, adult spotted seals on ice were characterized [52] as having low disturbance tolerance and hence being difficult to approach. Also, the hunters were encouraged to start harvesting from areas with high spotted seal concentrations in order to fulfill their catch quota. This explains an early spike in spotted seal catches and then a rapid and constant decline in numbers of this species harvested (see Fig 3D).

Changes in CPUE annual mean from year to year may be linked to organizational aspects of harvest operations, changes in harvest techniques, and staffing problems. Starting in 1975, new ice- reinforced vessels replaced the schooners and this change required additional staffing and training of new crews. The Korsakovskaya oceanic fishery government enterprise became responsible for vessel and crew preparation for the harvest season and the Okhotskrybvod later reported a highly unsatisfactory level of performance: there was not enough warm clothes, food supplies, and safety and shooting training provided to the newly employed vessel crews. Lack of sufficient hunting experience in a number of hunting brigades resulted in a two- to three-fold decline in CPUE (most pronounced in ringed seal harvests, which composed the major portion of catches), substantial increase in numbers of wounded seals and even led to an accidental injury of one of the hunters in 1977. The total time spent on-effort at sea also decreased due to testing and essential maintenance of new skiffs as well as crew at-sea training which took up to two weeks. The technical and staffing problems were gradually overcome after 1978 which resulted in increasing trends in catches (see Fig 4). A steady increase in ribbon seal CPUE could be related to the population recovery after an earlier period of uncontrolled exploitation, whereas increase of bearded seal take was mostly driven by abandonment of the fur-focused harvest which revived interest in this species. The CPUE of spotted seal increased shortly after the introduction of HFVs and remained stable until the end of the commercial harvest.

Declining CPUE values in exploited populations might be a good indicator of negative changes in population size. Visual inspection of GAM outcomes did not provide clear evidence of the decline of this parameter over the study period. Temporary troughs in CPUEs are shown to be largely associated with harvest management.

Spatial variation of catches and seal distribution

Spatial variation in catches was largely associated with vessel initial positions and harvest effort allocation. However, vessel harvest locations revealed significant differences in the distribution of catches of different seal species, which supported inferences about actual seal distributions on spring ice.

The ringed seal stands apart from other ice-breeding species due to its ability to maintain a system of water access and breathing holes in thick fast ice throughout the ice-covered period (e.g. [53]). This allows ringed seals’ breeding range to extend farther from the ice edge zones to avoid competition with other species for resources, and provides more access to ice features (hummocks and ridges; [9]) and dense snow cover as a shelter for their new born pups (e.g. [54]). In the Sea of Okhotsk, ringed seals have been found occasionally breeding on drifting ice where they do not always have the opportunity to construct birth lairs [51].

Based on aerial survey maps published in various years, ringed seals are fairly evenly distributed throughout their range in the Sea of Okhotsk [55–59]. However, the seal herds were not exploited equally by the sealers, which could lead to unbalanced exploitation of the most productive areas [9]. In April, for instance, large densities of ringed seals were reported in the Shantar Sea, but ice conditions early in the season prevented sealing vessels from working in this area.

Harvest data for 1972–1994 demonstrate that the main concentrations of ringed seals in April-May were closely associated with the near-shore fast-ice zone. A well-defined band of high seal density occurred along the coast from Shelikhov Gulf in the North to the northern Shantar Sea waters and the northern part of Sakhalin Island in the South. Fedoseev [56] described the breeding herd distribution for ringed seals and suggested that the groups of breeding seals south of the line from Ayan to St. Iona Island tended to concentrate on the so-called Shantarskiy ice massif. The other groups of breeding herds, located northward of 56°N, formed molting aggregations in the Okhotsk-Tauyskiy region. Breeding ringed seals were observed in Shelikhov Gulf, near the eastern Sakhalin and in the northern part of Terpeniya Bay [9]. In May, seals moved to molting grounds closer to the shore: north from Terpeniya Bay and north-west, west and south from the central Okhotsk Sea [60]. Based on our data (Fig B,C in S1 File), ringed seal molting herds, similar to other species, were associated with the remaining ice fields in late spring—early summer and were exploited exclusively in the Shantar Sea later in July (Fig D in S1 File).

The spatial structure of the ribbon seal harvest corresponded well with this species’ distribution in the whelping and molting seasons. The ribbon seal is a typically pelagic species whose habitat is associated with zones of the continental slope. It is incapable of maintaining breathing holes in the ice and occurs primarily in the areas where vertical circulation of water masses leads to formation of large ice floes, which makes these areas rich in fish. According to Fedoseev [9], in the Sea of Okhotsk, ribbon seals were observed more often on ice floes located in the deep-water zone off the shelf adjacent to Terpeniya Bay. Fedoseev et al. [57–59] survey maps showed ribbon seal aggregations in April being a considerable distance from shore, and also somewhat interior to the ice, away from the edge. In years with low ice concentration, ribbon seal herds could be found in the ice-edge zone beginning early in the season [56, 61], but in high ice concentration years, high densities of ribbon seals in April were only observed off the East coast of Sakhalin [62]. Comparing harvest distribution maps for April-early May (Fig E,F in S1 File) with the results of aerial surveys carried out in peak reproductive season [56–59] there were similarities in locations of high CPUE values and concentrations of ribbon seals. In both cases there were hot spots south of Terpeniya Bay in April, shifting gradually to the waters along the northeastern coast of Sakhalin Island later in the season if the ice conditions were initially severe or occurring there already in April if the ice edge was closer to the shoreline. Those two areas of high ribbon seal densities are isolated from one another by a vast zone of compact ice fields [9], which is not a preferred habitat for the species.

A dense concentration of ribbon seals was typically located north of 58°N and was considered as a separate breeding herd isolated presumably by bathygenic inflow of the warm Pacific waters passing through the Kashevarov Bank towards St. Iona Island [9]. Based on sealing log book data this herd was exploited only in May with occasional individuals taken in April. The distribution of molting ribbon seals in May reported by Fedoseev [56] coincided with the retreating ice edge location: seals were concentrated around the Shantar Sea and the north-western parts of the Sea of Okhotsk (Fig F in S1 File).

The bearded seal is a typical benthic and epibenthic feeder that prefers shallow, continental shelf regions. In wintertime it breeds on grey, grey-white, and white ice with a thickness of typically less than 60 cm [9] that prevails in the regions adjoining the fast ice polynyas and occurs very often over the continental shelf. In the Sea of Okhotsk, bearded seals prefer to breed closer to the shore and the distribution of breeding herds is rather stable from year to year, though in some places small patches of breeding seals are scattered chaotically throughout the shelf zone [9]. Bearded seals are not dependent on stable sea ice since bearded seal pups undergo prenatal molt and are capable of entering the water soon after birth [49]. Bearded seals in the study area breed later in the season than other seal species with peak pupping in late-April to mid-May [3, 63, 64].

Compared to ringed seals, bearded seals are less active in making and maintaining breathing holes in ice and require presence of water leads and polynyas. Thus, their breeding habitat in the Sea of Okhotsk is mostly located in the drifting ice zones in Shelikhov Gulf, Tauyskaya Bay and off eastern Sakhalin [56]. Their breeding habits correspond with catch distributions in April known from sealing log books (Fig I in S1 File). The tendency of bearded seal breeding and molting herds to associate with water openings likely leads to a patchy pattern of seal distribution throughout the sea, which is illustrated by aerial survey maps [56] and is supported by harvest distributions (Fig J in S1 File).

Bearded seals in their molting season remain patchily distributed [58], but the overall extent of the species coincides with the residual ice area. Maximal catches of the species in June and July occurred in the Shantar Sea coastal waters (Fig K,L in S1 File).

Spotted seal breeding areas are very similar to the ones occupied by ribbon seals. Fedoseev [9] suspected some degree of interspecific competition for available breeding habitats and fish resources. He also described spotted seal breeding herds as occupying the northern-northwestern part of the sea. This is under the influence of warm Pacific water masses penetrating through a trench along western Kamchatka, causing broken forms of ice to prevail [9]. Our data suggest the similar distribution of major reproductive concentrations of this species in April, and also confirms the presence of large numbers of spotted seals in Terpeniya Bay (Fig M in S1 File).

There are two major molting concentrations of spotted seals [56], but mainly the northern one was exploited during the harvest in May (Fig N in S1 File), which might be explained by later weaning of spotted seal pups in the northern parts of the Sea of Okhotsk [46], and hence more successful harvests in those areas due to higher disturbance tolerance in breeding females.