As a component of a complex program studying musk deer ecology in the Sikhote-Alin Biosphere Reserve, a survey of the musk deer population density, based on pellet group count method, was carried out at stationary posts from 2012 to 2014. The use of this method in the snowless season provides a means of surveying the most hard-to-reach areas of the musk deer habitat.
Musk deer ; Moschus moschiferus ; Population size survey ; Population density ; Pellet group count ; Sikhote-Alin Biosphere Reserve
Musk deer is a game animal in most parts of musk deer-inhabited areas in Russia. It is hunted mainly for the sake of preputial gland secretions; the musk is used in perfumery and oriental medicine. Demand for the musk is very high. According to a study conducted by the World Wide Fund for Nature (WWF), the volume of illegally traded musk amounted to 400–450 kg from 1998 to 2001 (Homes, 2004 ). Intensive unregulated hunting, which has thrived since the early 1990s, has led to a significant decrease in the number of musk deer in most parts of musk deer-inhabited areas in Russia. By the estimate of Centrokhotcontrol (Central Hunting Control) (Bersenev et al., 2011 ), in the mid-1980s, the musk deer population was 5–6 times greater than it is today. From 2008 to 2010, the number of musk deer in Russia was estimated to be 130–140 thousand individuals (Bersenev et al., 2011 ). In Primorskiy Krai, the population of the animal, resulting from winter route surveys from 2003 to 2010, was represented with a variety of estimates; the highest value was registered in 2003, with 17.43 thousand individuals (Bersenev et al., 2011 ). However, in other published estimates (Homes, 2004 ; Prikhodko, 1997 ) the quantity of the musk deer in Primorskiy Krai and Far East Russia was significantly different from Centrokhotcontrols data. The decrease of the musk deer population has been observed since the late-1970s in Central Sikhote-Alin as well, including the Sikhote-Alin Biosphere Reserve.
Estimating the size of the musk deer population is a significant challenge. There is a widespread belief that there are no adequately developed survey methods (Homes, 2004 ). Nevertheless, there are several methods of surveying musk deer (Zaitsev et al., 2013 ). Differences in the estimates of musk deer abundance are caused by the use of different methods. The use of the winter route counting method, which is popular in Russia, is associated with a number of problems for estimating the musk deer population density (Zaitsev, 1991 ; Zaitsev, 2006b ). The data obtained with this method show the overall population dynamics of musk deer, but usually distort the estimates of population density. However, there are other survey methods that have been developed in the Sikhote-Alin Biosphere Reserve, including surveys based on pellet group count method in snowless seasons. The use of this method on sufficiently long survey routes allows an estimation of the density of the musk deer population with an accuracy of no more than 6–10%, with the indicators taken into account in winter test surveys using other methods (Zaitsev, 2006b ; Zaitsev et al ., 2013 ).
The method of musk deer surveying based on pellet group count has advantages over other methods (Zaitsev, 2006b ; Zaitsev et al ., 2013 ). The advantages of this method include, primarily, more favorable conditions for the surveyors, relative ease of movement during frost-free and snowless seasons, and ability to estimate musk deer population density in remote and inaccessible places that are favored by musk deer.
A musk deer abundance survey was carried out in 2012 within the framework of the program for the study of the ecology of this species (Maksimova et al., 2014 ) on the Sikhote-Alin State Nature Biosphere Reserve in the basins of the Tayozhnaya (upstream), Dzhigitovka (Nevidimka stream) and Serebryanka (Zimoveyniy stream) Rivers and in 2014 in the basins of the Tayozhnaya (upstream) and Serebryanka (Zimoveyniy stream) Rivers. In 2012 and 2014, the survey that was carried out in the basin of the Tayozhnaya River was also carried out at the territory adjacent to the reserve (Sedmoy Creek).
All of these areas are characterized by abundant mountain pine forests with various additions of broadleaved and dark coniferous (spruce, fir) species, both in the stand and in young growth. These sites are the typical habitat of musk deer in Sikhote-Alin (Zaitsev, 1991 ; Zaitsev, 2006a ; Zaitsev, 2006b ). However, pine forests outside the Reserve have been subjected to anthropogenic influences for a long period of time (e.g., forestry and hunting).
To estimate the population density of musk deer, the survey based on pellet group count method was carried out in the snowless season (Zaitsev, 2006b ; Zaitsev et al ., 2013 ). The survey routes were plotted at random, crossing different habitats of musk deer. While moving along the route, the surveyors searched on both sides and marked all musk deer pellet groups. The route was recorded on a GPS-navigator system. The survey was conducted from May until mid-June when there was no snow and the grass cover was not thick enough to complicate the search for pellet groups.
All of the latrines along the route (places where musk deer of different sexes and ages leave their excrements, see Fig. 1 ) and separate pellet groups located no closer than 0.5–1 m from the latrines were considered the unit of measurement (Zaitsev et al., 2013 ). The condition of the pellets varies in different places depending on many factors (e.g., forest type, vegetation development, moisture, latrines refreshing). Therefore, the figures obtained with this method do not show the current musk deer population density, but rather a retrospective density that exists mainly from the autumn until the spring (Zaitsev et al., 2013 ).
Latrine of musk deer.
To estimate the population density of musk deer, we used the method based on the standard ratio of the number of musk deer pellet groups counts along the route during the snowless period and the population density that was calculated at the survey sites in the winter using other methods. The excrements decomposition rate varies in different habitats. Therefore, the probability of detecting pellets has various limitations in different types of forests. For this reason, this method suggests the use of several different formulas depending on the conditions that were present when the work was carried out (Zaitsev, 2006b ).
The survey routes were located away from lengthy trails of ungulates and only crossed small fragments of larch forests. Therefore, when determining the population density of musk deer, we applied the following formula:
where P is musk deer population density per 10 km2 and N is the average number of latrines and pellet groups per 1 km of the route.
In addition to the absolute density, we calculated the relative density of musk deer, as expressed in the number of latrines and pellet groups, per 1 km of the route.
The total length of the survey routes in both 2012 and 2014 was 159.73 km (72.25 km in 2012 and 87.48 km in 2014). A total of 242 latrines and pellet groups of musk deer were found (120 and 122 in 2012 and 2014, respectively).
In the surveyed areas, the greatest population density of musk deer was recorded in 2012 in the basin of the Tayozhnaya River in the Sikhote-Alin Biosphere Reserve. Here, 3.03 pellet groups per 1 km of the route and 4.6 individuals per 10 km2 (see Table 1 ) were observed. A somewhat lower population density was registered in the same area in 2014, 2.51 pellet groups per 1 km of the route and 3.8 individuals per 10 km2 . At the adjacent site located in the basin of the same river but outside of the reserve, the musk deer population density was significantly lower than that in the Reserve, both in 2012 and 2014, and did not exceed 1.37 pellet groups per 1 km of route and 2.2 individuals per 10 km2 (see Table 1 ).
|Year, river basin, territory status||Route length, km||Number of latrines and pellet groups along the route||Relative density of pellet groups per 1 km of route||Population density, individuals per 10 km2|
|2012, Tayozhnaya River, Reserve||28.71||87||3.03||4.61|
|2012, Tayozhnaya River, adjacent territory||16.73||23||1.37||2.19|
|2012, Dzhigitovka River, Reserve||17.18||8||0.47||1.12|
|2012, Serebryanka River, Reserve||9.63||2||0.21||0.86|
|2014, Tayozhnaya River, Reserve||42.18||106||2.51||3.78|
|2014, Tayozhnaya River, adjacent territory||25.98||16||0.62||1.29|
|2014, Serebryanka River, Reserve||19.32||0|
The data presented in the table also indicate an increase in the overall density of musk deer population from the southern parts of the Reserve (basin of Dzhigitovka River, Reserve), where pine forests include a significant proportion of broadleaf trees, in contrast to the northern areas (Tayozhnaya River, Reserve), where conifers (especially fir) are more represented by forest stands and young growth. These features of the musk deer distribution were also observed earlier, during the period of its high population (Zaitsev, 1991 ).
The lowest values of the parameter under study were registered in 2014 at the Zimoveyniy site in the basin of the Serebryanka River, when not a single pile of musk deer pellet groups was found along the 19.32-km survey route (see Table 1 ). A decrease of the musk deer population in this area, as well as throughout the river basin, has been observed since the 1980s (Zaitsev, 1991 ; Zaitsev, 2006b ). From 1975 to 1981, the density of the animal group that was traced by us at the Zimoveyniy site reached 11–19 individuals/10 km2 . In the period from 2008 to 2010, according to winter surveys, there was a particularly sharp decrease in the musk deer population density in the survey area, to the extent of its complete disappearance in valleys and slopes adjacent to the river.
The use of a complex technique (tracking, radio-registration, camera traps and special survey methods) within the framework of the program for studying the ecology of musk deer (Maksimova et al., 2014 ) confirms the feasibility of musk deer population density data obtained in the course of the survey based on pellet group count, which was conducted during a period of decreased population of the species.
In comparison with the results of surveys conducted from the 1970s to the 1980s, the current survey data show a profound decline in the numbers of musk deer in the Reserve and especially in the adjacent territories.
The comparison of the musk deer population density obtained by the survey based on pellet group count in specially protected natural areas and adjacent territories that are subject to human impact leads to the conclusion that in the absence of human impact or with minimal impact on the population of musk deer in the reserve, its density is much higher than that in areas with intensive hunting and logging. These results are consistent with those obtained by other methods. It was shown earlier, in an example of the same area, that the probability of the presence of musk deer is greatly affected by logging (Zaitsev, 2006b ; Slaght et al ., 2012 ).
According to the data obtained, the musk deer population density in the Reserve is currently at a low level. In comparison with a high population period (from the 1970s until the 1980s), the data from surveys conducted in 2012 and 2014 indicate a multifold (sometimes more than 10-fold) decrease in the density of the population in many of the areas under study where previously it had been 10–20 individuals per 10 km2 (Zaitsev, 2006a ).
In adjacent areas, the principal reason for the decrease in the musk deer population is undoubtedly unregulated, and often illegal, hunting. However, the reasons for the decline and fluctuations in the numbers of musk deer in the Reserve have not been studied sufficiently, and monitoring the population density at fixed locations is necessary.
A relatively convenient and simple method for monitoring the condition of populations (groups) of musk deer is estimating the relative and absolute density of the population on survey routes based on pellet group count during snowless time periods. The application of this method can be recommended to estimate the abundance of musk deer in specially protected territories as well as in hunting areas.
The authors would like to thank N.N. Rybin, V.V. Melnikov (Wildlife Conservation Society), S.V. Soutyrina, I.M. Mogilnikova (Sikhote-Alin Biosphere Reserve), Yu.K. Petrunenko, M.Yu. Borisov (Pacific Geographical Institute FEB RAS), E.A. Petrunenko (Botanical Garden-Institute FEB RAS) and V.M. Shaitanov (K.I. Skryabin Institute of Fundamental and Applied Parasitology of animals and plants) for assistance in the field studies.