Popul Ecol (2010) 52:373–387 DOI 10.1007/s10144-010-0219-4
SPECIAL FEATURE: REVIEW
Adaptive Management
Adaptive management of sika deer populations in Hokkaido, Japan: theory and practice Koichi Kaji • Takashi Saitoh • Hiroyuki Uno Hiroyuki Matsuda • Kohji Yamamura
•
Received: 5 March 2010 / Accepted: 9 May 2010 / Published online: 1 June 2010 Ó The Society of Population Ecology and Springer 2010
H. Matsuda Department of Environmental Management, Yokohama National University, Yokohama 239-8501, Japan
hunting pressure were configured, with a choice of four corresponding management actions. Under this program, the Hokkaido Government has been promoting aggressive female culling to reduce the sika deer population since 1998. We devised a harvest-based estimation for population size using relative population size and the number of deer harvested, and found that the 1993 population size (originally estimated by extrapolation of aerial surveys) had been underestimated. To reduce observation errors, a harvest-based Bayesian estimation was developed and the 1993 population estimate was again revised. Analyses of population trends and harvest data demonstrate that hunting is an important large-scale experiment to obtain reliable estimation of population size. A serious side effect of hunting on sika deer was inadvertent lead poisoning of large birds of prey. The prohibition of the use of lead bullets by the Hokkaido Government was successful in reducing the lead poisoning, but the problem still remains. Two case studies on sika population irruption show that the densities set by maximum sustainable yield may be too high to prevent damage to agriculture, forestry, and/or ecosystems. Threshold management based on feedback control is better for ecosystem management. Since volunteer hunters favor higher hunting efficiency in resource management (e.g., venison), it is necessary to support the development of professional hunters for culling operations for ecosystem management, where lower densities of deer should be set for target areas. Hunting as resource management and culling for ecosystem management should be synergistically combined under AM.
K. Yamamura Laboratory of Population Ecology, National Institute for Agro-Environmental Sciences, Tsukuba 305-8604, Japan
Keywords Cervus nippon Density dependence Feedback management Maximum sustainable yield Population dynamics
Abstract We investigated the utility of adaptive management (AM) in wildlife management, reviewing our experiences in applying AM to overabundant sika deer (Cervus nippon) populations in Hokkaido, Japan. The management goals of our program were: (1) to maintain the population at moderate density levels preventing population irruption, (2) to reduce damage to crops and forests, and (3) to sustain a moderate yield of hunting without endangering the population. Because of significant uncertainty in biological and environmental parameters, we designed a ‘‘feedback’’ management program based on controlling hunting pressure. Three threshold levels of relative population size and four levels of
K. Kaji (&) Faculty of Agriculture, Tokyo University of Agriculture and Technology, 3-5-8 Saiwaicho, Fuchu 183-8509, Japan e-mail:
[email protected] T. Saitoh Field Science Center, Hokkaido University, Sapporo 060-0811, Japan H. Uno Institute of Environmental Sciences, Hokkaido Research Organization, Sapporo 060-0819, Japan
123
374
Popul Ecol (2010) 52:373–387
Adaptive management (AM) is a management system for natural resources through continually improving management policies and practices, including both approaches of adaptive learning and feedback control (Walters 1986). Feedback management or feedback control is a management policy for sustainable use of natural resources at an optimal level under uncertain information of resources, which was proposed by Tanaka (1982). Taking in its broad sense, feedback management is almost the same as AM. To put it concisely, AM is a scientific way of ‘‘learning by doing’’, changing management tactics according to the results of management action and improving a program when new knowledge is gained and/or an error of past knowledge is found (Christensen et al. 1996). It could work efficiently to manage populations or resources with uncertainties under insufficient information. Although the essence of adaptive management is rather simple, it requires many efforts to accomplish its processes: setting the objectives of the management, estimating ecological parameters, planning the management based on the predicted dynamics, and monitoring of the results. Often, specific survey protocols and new statistical methods for estimation and prediction are needed. Although the AM approach has been considered the best way for resolving both the biological and political dilemmas surrounding deer management in the US National Parks (Porter and Underwood 1999), applying AM to deer management has been very limited. A sort of adaptive management is employed in France for managing the roe deer (Capreolus capreolus) using indicators of ecological change to monitor populations at a local scale (Morellet et al. 2007). The AM approach, however, has never been applied at a wide spatial scale, because the indicators of ecological change have been validated only in well-defined populations at a small scale (Morellet et al. 2007). 90,000
males (nuisance control)
80,000
males (hunting)
6,000 5,000
females (nuisance control)
70,000
Number of harvest
Fig. 1 Changes in sika deer (Cervus nippon) harvest and agriculture/forestry damage occurring between 1957 and 2008 in Hokkaido (Hokkaido Government, unpublished data)
Hokkaido is the northernmost island of Japan and covers 78,073 km2. It was relatively recently (the late nineteenth century) industrialized by the Japanese Government. Sika deer (Cervus nippon) populations in Hokkaido are characterized by a relatively simple life history (Hokkaido Institute of Environmental Sciences 1997). The mean life span of males (2.1–3.1 years) was shorter than for females (3.6–3.9 years). First ovulation and/or pregnancy occurs during the breading season at yearling age. Pregnancy rate of adult female sika deer (2 years and older) is higher than 90%, which is maintained throughout their life. Litter size is typically one and twinning is rare. The history of sika deer populations in Hokkaido is characterized by overexploitation and protection (see ‘‘Historical background of sika deer management’’ for details). The overharvest by commercial hunting in the late 1800s and a heavy snow in 1879 resulted in a dramatic decline in deer abundance, while the distribution range shrank into limited areas. Due to long-term legislated protection such as bans on hunting (1890–1900 and 1920– 1952), and bucks-only hunting from 1955, the sika deer population gradually recovered its distribution, becoming well established in eastern Hokkaido by the mid-1970s, and had spread to potential habitats all over Hokkaido by the 1990s (Kaji et al. 2000). Abundance has also greatly recovered, causing severe damage in agricultural and forestry during the last three decades (Fig. 1). In 1998, in response to this damage, the Government of Hokkaido implemented the Conservation and Management Plan for Sika deer (CMPS) in eastern Hokkaido and promoted aggressive population control based on AM (Hokkaido Government 1998; Matsuda et al. 1999). During preparation for the CMPS, we faced a lot of uncertainties in estimating the population size and demographic parameters because of insufficient information. Because of these large uncertainties, we have been exploring how to apply AM to deer management practices since 1998. This paper will
females (hunting)
4,000
60,000 damage
50,000 3,000 40,000 2,000
30,000 20,000
1,000 10,000
Year
123
05 20
00 20
95 19
90 19
19
85
80 19
75 19
70 19
65 19
60
0
19
19
55
0
Damage incurred (million yen)
Introduction
Popul Ecol (2010) 52:373–387
show the usefulness of AM in wildlife management, reviewing our experiences in applying AM to the sika deer population in Hokkaido. In this review paper, we first describe the historical background of sika deer management in Hokkaido since the late nineteenth century, and then introduce theory and practices of AM for sika deer populations in Hokkaido. Through the practices, we have learned the importance of the following points: (1) monitoring systems to recognize dynamics of deer populations, conflict with human activities, and their impacts on ecosystems; (2) statistical techniques to estimate ecological parameters and to predict dynamics of deer populations for making a management program; and (3) adaptive learning to solve unpredictable problems that were brought about by management treatments. Finally, based on our long-term study of two sika deer populations, we argue a problem with maximum sustainable yield (MSY) theory for the application to sika deer management concerning the protection of vegetation. In ‘‘Conclusion: hunting as resource management and culling for ecosystem management’’, we emphasis the importance of hunting as resource management and of culling for ecosystem management.
Historical background of sika deer management Overexploitation in the second half of the nineteenth century resulted in major declines in both deer numbers and range worldwide. Especially in Europe and North America, subsequent protection of deer (hunting regulation and reduction of natural predators), habitat changes due to agricultural and silvicultural activities and moderate climates then caused rapid population increases (McShea et al. 1997; Linnell et al. 1998; Coˆte´ et al. 2004). The history of sika deer populations in Hokkaido (Table 1) is in some respects similar to that of deer populations in North America. Prior to the Japanese colonization of Hokkaido in the late nineteenth century, sika deer had occurred throughout Hokkaido. Several factors such as overexploitation, habitat loss due to agricultural development and timber extraction in lowland forests, and heavy snowfalls contributed to the rapid decline in deer numbers around 1900 (Inukai 1952). Between 1873 and 1878, the annual harvest ranged from 60,938 to 110,002 deer. In 1876, in an effort to control hunting, the Hokkaido Government established a modern hunting law (Table 1; Tawara 1979), while establishing a deer meat cannery and encouraging exports of deer products (meat, hide and antlers) in 1878 (Inukai 1952). The new hunting regulations were based on the recommendations of Horace Capron, a technical advisor from the US, who was familiar with the legislation proposed to prevent the extirpation of
375
the North American bison (Bison bison). However, regulation was not able to prevent a drastic decrease of the sika deer populations; overharvesting and heavy snow in 1879 led to a great winter mortality, and it is thought that populations declined to a threatened level. Three major populations survived this bottleneck, in the Akan, Hidaka, and Daisetsu mountain regions (Fig. 2). To conserve the sika deer populations, hunting was banned twice, in 1890–1900 and 1920–1952 (Table 1). During the Second World War, deer populations gradually recovered, such that bucks-only hunting was allowed in parts of the Hidaka district in 1955–1956, and the hunting areas expanded more widely in 1957. Extirpation of wolves by 1890 (Inukai 1933), hunting regulation (prohibition of hunting and bucks-only hunting), replacement of native mixed-hardwood forests with conifer plantations, and increased pasture acreage may have contributed to the recovery of deer distribution and abundance (Kaneko et al. 1998). By the mid-1970s, sika deer occupied most available habitats in the eastern half of Hokkaido and had spread to the south and northwestern parts of the island and had occupied their entire potential habitats by the 1990s (Kaji et al. 2000). Legislated protection such as bucks-only hunting and large game reserves (required, when setting hunting areas, to cover areas being equivalent to 1/3 of the hunting area) greatly contributed to the recovery of the populations and to the maintenance of the hunting system. However, the success of the protection measures turned out to be a cause of overabundance. As the range of sika deer expanded, agricultural and forest damage increased. Damage to agricultural crops and forests by sika deer remained at low levels from the mid-1950s through the mid-1970s, but had increased dramatically to nearly 2 billion Japanese yen (JPY) by 1990 and to over 5 billion JPY by 1996 (Fig. 1). In Nature Reserves, such as the Akan National Park and Shiretoko National Park, elm trees (Ulmus laciniata and U. davidiana) were seriously debarked by deer and the increase in standing dead trees became obvious. Road kills and rail accidents also increased and human injury caused by road accidents, animal welfare (e.g., road kills) and economic impacts became serious. In response to the damage, hunting regulations were relaxed, but female hunting was not allowed until 1994 (Table 1; Fig. 1). The total harvest fluctuated between 2,000 and 3,000 animals during 1970–1989 and increased to 16,134 animals in 1990 and to 46,634 animals in 1996. One of causes of the problem was that no goal was ever set for the protection of the deer. In 1990, experts of bear and deer biology (who took part in the 5th International Ecological Congress held in Yokohama, Japan) from European and North American countries were invited to Hokkaido. Through the ‘‘Deer
123
376
Popul Ecol (2010) 52:373–387
Table 1 Status of sika deer (Cervus nippon) population and government control on hunting in Hokkaido Year
Status of deer population
1873
Population begun to decrease
Government control on hunting Extensive market hunting
1876
The number of hunters, period and season were regulated and the use of aconitum-tipped poison arrows was prohibited
1878 1879
A deer meat cannery was established Drastic decrease by overharvest and heavy snow
1890–1900
Near extinction
Ban on hunting
1920–1952
Threatened level
Ban on hunting
1955
Gradually recovered
Bucks-only hunting started
1991
Gradually increased
Monitoring started by HIES
1994–1996
Damage income of agriculture and forestry by deer reached 5 billion JPY in 1996
Female hunting started in restricted areas (8–10 municipalities) for 10 days each year. Allowable harvest was 1 deer (1 male or 1 female) per day
1997
Female hunting areas extended to 63 municipalities and the hunting period to 30 days
1998
Population reached a peak in eastern HK
Aggressive female harvesting started based on CMPS. Allowable harvest increased to 2 deer (2 females or 1 male ? 1 female) per day, and the hunting season extended to 3 months
2000
Population decreased in eastern HK, while increased in western and northern HK from 1998 to 2000
Management area extended to include central part of HK. Hunting seasons was extended to 4 months.
2001
Population increased in eastern HK
Banned the use of lead bullets in shot guns. Allowable harvest extended to 3 deer (3 females or 1 male ? 2 females)
2002
Population increased all over HK
2004 2005 2008
Banned the use of lead bullets in rifles
Management areas extended all over HK Unlimited female hunting started
Population recovered to almost the same level as 1993 in eastern HK Population irrupted all over HK
Encouraging sustainable resource management of deer as venison
JPY Japanese yen, HK Hokkaido, HIES Hokkaido Institute of Environmental Sciences
and Bear Forum Hokkaido 1990’’ held in Sapporo, they proposed a modern wildlife management system for the Hokkaido Government referring to those in their countries (Ohtaishi et al. 1990). They recommended that the Hokkaido Government establish a wildlife management agency and promote integrated management systems for wildlife. The Hokkaido Government responded promptly to their proposal and established the Hokkaido Institute of Environmental Sciences (HIES) in 1991. Monitoring of sika deer populations was started using population indices such as spotlight counts, aerial surveys, catch per unit effort (CPUE; number of deer harvested per hunter-day), sighting per unit effort (SPUE; number of deer sighted per hunterday), and damage to agriculture and forestry. This was the beginning of a more scientific approach to managing sika deer populations in Japan. Intensive studies on seasonal migration, age structure and reproductive condition were also begun on model study sites. In response to the increase in damage by sika deer, female hunting was allowed from
123
1994 in restricted areas and periods. Due to the strong social pressure against female hunting from citizens and even from some wildlife biologists who were concerned about a marked decrease of sika deer populations, we were not able to promote aggressive female hunting at this stage. Although the Hokkaido Government extended female hunting areas to 63 municipalities in 1997 (about 30% of Hokkaido as of 1997) and the hunting period (increasing it by about 20-fold compared to those in 1994–1996; Table 1), the number of harvested females in this year increased only by a factor of 1.4 under the hunting regulation of one female or one male per day. This experience revealed that hunters still favored males, and thus a sexspecific hunting regulation was the key to increasing the female harvest (Matsuda et al. 1999). Lesson 1. The success of one management action could turn out to be the cause of another problem (e.g., overabundance). Clear goals for management programs and monitoring the effect of the program are essential to avoid generating new problems.
Popul Ecol (2010) 52:373–387
Daisetsu
Hidaka
377
Akan
1925
1954
1974
1991 1980: 6 Females, 2 Males 1981: 6 Females, 3 Males
1997 (supplementary survey in 2005)
Fig. 2 Sika deer range expansion from 1925 to 2005 in Hokkaido, Japan, estimated from observations reported in personal interviews and mail surveys. Arrows at the southern end of Oshima Peninsula indicate sites where deer were released in 1980 and 1981. Deer populations expanded from the Hidaka, Daisetsu and Akan mountain regions. The historical distribution map was modified from Kaji et al. (2000, 2006)
Theory and monitoring for adaptive management of sika deer in Hokkaido Spatial distribution, population structure and monitoring unit To obtain information about deer distribution, the Hokkaido Government conducted personal interviews or systematic mail surveys in all municipalities of Hokkaido in 1978, 1984, 1991, and 1997 with a supplementary survey in 2005 (Kaji et al. 2000, 2006; Fig. 2). These surveys revealed the recovery process of deer from threatened levels around 1900. At least three populations of sika deer survived the bottleneck and expanded from a few natural refuges in the mountainous areas of Akan, Hidaka, and Daisetsu districts (Nagata et al. 1998; Kaji et al. 2000, 2006; Fig. 2).
Logistic regression analyses of the effect of environmental conditions on deer occurrence based on GIS data suggested that snow depth and bamboo grass (Sasa spp.) variation were important in the regulation of sika deer distribution, and that deer had occupied almost all the potential range by 1990 (Kaji et al. 2000). Part of the recent expansion resulted from experimental reintroductions in 1980 (6 females and 2 males) and in 1981 (6 females and 3 males) on Oshima, Hokkaido’s southernmost peninsula (Fig. 2; Kaji et al. 2000). The distribution ranges have been further expanding in western and northern Hokkaido (Fig. 2) where there is normally heavy snow accumulation. This range expansion may be caused by recent warm winter climate conditions. Based on these distribution maps, we set up 12 monitoring units in Hokkaido (Fig. 3) and assessed population status in each (Hokkaido Institute of Environmental Sciences 1994). These ‘‘monitoring units’’ were not purely populationbased, but also relied on both geographical features and District office boundaries. Each District office is composed of several ‘‘municipalities’’. Since most sika deer in eastern Hokkaido make large-scale seasonal migrations (Sakuragi et al. 2003; Igota et al. 2004), 4 monitoring units (units 9–12) located in eastern Hokkaido were combined as the area for management of the target population (the Akan population; Fig. 3). The Akan population was the largest, and caused the most serious damage associated with rapid range expansion and population increase during the 1980s and 1990s, so we focused on management of the Akan population. The main wintering grounds of the Akan population were the Akan National Park and neighboring Shiranuka Hills (Uno et al. 2009). Deer showed strong site fidelity to their seasonal ranges (Uno and Kaji 2000; Igota et al. 2004). The summering areas are extensive and contiguous in eastern Hokkaido (Kaneko et al. 1998), while potential wintering areas are restricted to a few limited areas because of heavy snow and availability of coniferous forests for shelter (Kaneko et al. 1998; Sakuragi et al. 2003). Of 57 radio-collared female deer captured in the Shiranuka Hills (a wintering site), 21% were non-migrants and 69% were migrants (Igota et al. 2004). Monitoring by population indices Since complete counts across an entire sampling frame or study area are rarely possible, index methods using sample counts are a practical way to monitor the change in the size of a target population and to evaluate the estimation error (Thompson et al. 1998). There are, however, many problems with index methods that result in a lack of rigor and validity (White 2001). We compared the rigor and validity of the following population size indices; spotlight counts
123
378
Popul Ecol (2010) 52:373–387 N
300
Eastern Hokkaido
12
9
6
Population index
7
3
Western Hokkaido
250
8
200
150
100
11 5
2
50
0
123
males (hunting) males (nuisance control)
160
70,000
140
60,000
120
50,000
100
40,000
80
30,000
60
20,000
40
10,000
20
00 20 01 20 02 20 03 20 04 20 05
99
20
98
19
97
19
96
0
19
94
19
19
0
Population Index
females (hunting) females (nuisance control)
80,000
95
on farmland, aerial survey on wintering ground from a helicopter, CPUE, SPUE, and the value of damage to agriculture and forestry (Uno et al. 2006). The index of aerial surveys was available only for specific ranges (Akan National Park and Shiranuka Hills) and limited years (February–March 1993, 1994, 1997–2002). The index based on the aerial surveys did not show a clear correlation with any other indices. Hunters’ reports and the resulting CPUE index were affected by changes in hunting regulations. While, on the other hand, changes in the hunting regulation do not seem to have affected the SPUE index, the SPUE index varied among years and did not show a consistent trend. In addition, there was a 2-year time lag until the SPUE index became available, because data processing was time consuming work. Thus, the SPUE index was inferior to the spotlight count index which had a shorter time lag of 1 year, although the spotlight count index showed large annual variation. The index based on damage by deer is influenced by agricultural practice and has been decreasing recently due to the construction of deer-proof fences. The spotlight survey index (Fig. 4) appeared to track estimated population changes (Figs. 5 and 6); it increased from 1992 to 1998, and decreased thereafter, which coincided with a period of aggressive
Fig. 4 Population index estimated by spotlight count in eastern and western Hokkaido. The indices were standardized by the value of 1993 in eastern Hokkaido, and 2000 in western Hokkaido (Hokkaido Government, unpublished data)
19
Fig. 3 Monitoring units for sika deer in Hokkaido. Each unit is identified by a number (Hokkaido Institute of Environmental Sciences 1994). The shaded area indicates the management area for the Akan population (monitoring units 9–12, comprising 40 municipalities, the dark shaded area was where the aerial survey was conducted) In dark shaded and middling shaded areas, we could use the spotlight counts, CPUE and SPUE data (25 municipalities, two spotlight counts data were excluded for analysis because of net fences). In light shaded areas, we could use only spotlight count data (15 municipalities). We did aerial surveys in the dark shaded area (3 municipalities) (Uno et al. 2006)
Year
93
200 km
19
100
Number of harvest
0
94
19
19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08
4
19 93
1
10
Year
Fig. 5 Harvest-based estimates of population indices (Matsuda et al. 2002) and number of deer harvested from 1993 to 2005 in eastern Hokkaido. Bold line the estimates of population index, dotted lines SE
EPC culling during which more than 27,000 female deer were killed annually (Figs. 5 and 6). In addition, a reconstructed sika population by cohort analysis showed a consistent trend with the population index by spotlight survey (Ueno et al. 2010). Furthermore, its estimate error is relatively small (Uno et al. 2006). In conclusion, the index based on spotlight surveys is the most useful of the five indices in looking at the population trend (Uno et al. 2006). However, spotlight surveys did show unrealistic low abundance in some specific years (e.g., in 1994), and thus the validity of the spotlight count index should be checked by using another index (e.g., SPUE).
Popul Ecol (2010) 52:373–387
379
females (hunting) females (nuisance control)
males (hunting) males (nuisance control)
a Eastern Hokkaido
100,000
160
120 100
60,000
80 40,000
60
Population Index
Number of harvest
140 80,000
40 20,000 20 0
94
19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08
19
93
0
19
350
60,000
b Western Hokkaido Number of harvest
250
40,000
200 30,000 150 20,000
100
10,000
Population Index
300
50,000
50 0
96 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 19
95
19
4
9 19
19
19
93
0
Year
Fig. 6 Bayesian estimates of population indices that were obtained from the state-space model based on the stage-structured model (Yamamura et al. 2008) and number of deer harvested from 1993 to 2008. Bold line the estimates of population indices, dotted lines SE. Since spotlight surveys were not sufficient in number in western Hokkaido until 1999, the population indices were obtained using the data between 2000 and 2008 for relative abundance, with the value for 2000 as 100%
Feedback management based on population dynamics model in eastern Hokkaido Population dynamics model is essential for scientific management planning because we need to predict population dynamics, to determine harvest numbers, and to evaluate management implementation by comparing the prediction with the results. Sika deer populations in Hokkaido are characterized by (1) a high intrinsic rate of increase (rm = 0.15–0.19; Kaji et al. 2004), (2) no significant density effects on population growth until just before a population crash (Kaji et al. 1988, 2004), and (3) a relatively simple life history (see ‘‘Introduction’’). Sika deer populations have fluctuated greatly due to heavy snow, overharvest, and bans on hunting in Hokkaido (Inukai 1952). Because of the large number of uncertain biological, environmental, and game-hunting parameters, the Hokkaido Government designed a ‘‘feedback’’ management
program which is the control of hunting pressure based on the prediction of the dynamical model. In 1998, the Hokkaido Government implemented the CMPS in eastern Hokkaido. In the CMPS, population indices were obtained from spotlight counts, hunting statistics, and the extent of damage of crops and forests that had been monitored every year. In fiscal year 1993, the population size in eastern Hokkaido was estimated at 120,000 ± 46,000 (90% CI) deer as of March 1994 based on aerial surveys (Hokkaido Institute of Environmental Sciences 1995). However, since its accuracy was uncertain, we used a relative population size index where the value for 1993 was 100%. Population control without setting a goal is like setting off on a voyage without a compass, so the management goals were carefully identified: (1) to maintain moderate population levels to prevent irruptive behavior of the populations, (2) to reduce damages to crops and forests, and (3) to sustain a moderate yield without endangering the population. Dynamics of the target population were predicted by a simple stage-structured model, considering the effects of uncertain population parameters and severe winters. Management actions based on population indices estimated by the model analyses and the prediction have been implemented by controlling hunting pressure (Matsuda et al. 1999). A standard feedback program determines harvest rate by a dynamic equation (Tanaka 1982). If the estimate of population is larger (or smaller) than an optimal size, hunting pressure is increased (or decreased) in proportion to the difference between the estimate and the optimal size. Under this program, the dynamics of deer population and hunting pressure are similar to prey–predator dynamics in Lotka–Volterra systems (Lotka 1925; Volterra 1926). If the population has strong density effects, the feedback system has a stable equilibrium, although the population fluctuates in the process of approaching the equilibrium. Thus, standard feedback management is effective when the current population size is relatively close to the target level and/or the population has a strong density effect (Tanaka 1982). The current population size of sika deer, however, greatly exceeded the target level and no significant density effect on population growth rate was expected. Thus, a model analysis showed that the applying standard feedback management program to the population dynamic model for the target population generated large fluctuations at intervals of 100 years, and the target population size could never be realized (Matsuda et al. 1999). As an alternative, we developed a model involving three threshold levels of relative population size, four levels of hunting pressure, and a choice of one of four actions, based on the estimates of relative population size (Table 2; Fig. 7; Matsuda et al. 1999). The three thresholds were
123
380
Popul Ecol (2010) 52:373–387
Table 2 Four management programs based on relative population size index Program
Population size index
Action
Emergent population control (EPC)
%P? \ %P(t)
Maximum harvest of females
Gradually decreased action (GDA)
%P* \ %P(t) \ %P?
Hunting for males and females
Gradually increased action (GIA)
%P- \ %P(t) \ %P*
Bucks-only hunting
Ban on hunting (BH)
%P(t) \ %P1
Ban on hunting -
Population size index
%P(t) Population size index observed, %P the irruption level, %P* the optimal leve, %P the critical level (Matsuda et al. 1999)
%P(t)
EPC
%P +
GDA %P*
GIA
%P -
BH
Year Fig. 7 Schematic diagram illustrating a framework of feedback management for the sika deer population in eastern Hokkaido; %P(t) population size index observed; %P? the Irruption threshold, %P* the Optimal level, %P- the Critical threshold, EPC emergency population control, GDA gradually decreased action, GIA gradually increased action, BH ban on hunting (Hokkaido Government 1998)
determined using risk assessment under 4 scenarios (Matsuda et al. 1999): (1) a Critical threshold (%P-) where the probability [the number of adults is smaller than 1,000 within the next 100 years] must be smaller than 1%, (2) an Optimal level (%P*) where the probability [the estimate of population index (%P) is smaller than %P- in the next 100 years] must be smaller than 2.5%, and (3) an Irruption threshold (%P?) where the probability [the estimates of relative population index (%P) is larger than %P? in the next 100 years after %P is less than %P?] must be smaller than 2.5%. The four potential scenarios are as follows (Table 2). If the current deer population index [%P(t)] were above the irruption threshold (%P? = 50% of the population indices in 1993), emergency population control (EPC) would be adopted; i.e., maximizing the harvest of female deer by hunting and control kill. If %P(t) were larger than the optimal level (%P* = 25%) but smaller than %P?, hunting for males and females would be allowed by gradually decreased action (GDA). If %P(t) were larger than the critical threshold (%P- = 5%) but smaller than the optimal level (%P*), bucks-only hunting would be allowed by gradually increased action (GIA). If %P(t) were smaller than %P-, a total ban on hunting (BH) is implemented.
123
Sex-specific hunting is a key to maintaining a considerable amount of harvest. Hokkaido Government decided to adopt EPC in 1998– 1999; hunting area and period of hunting season for females were extended to sixfold and threefold larger than those in 1994–1996, respectively. Because of uncertainty of actual increment of hunting effort and population abundance, the Hokkaido Government relaxed hunting regulations to maximize harvest of female deer step by step according to the basic concept of feedback management (Table 1). Lesson 2. Since information about wildlife populations has uncertainties, investigators should monitor populations using several independent population indices, cross-check them, and confirm the consistency.
Practice of adaptive management: learning by doing, through emergency population control, 1998–2008 In 1998, since population indices exceeded the irruption threshold (Fig. 5), the Hokkaido Government implemented EPC (Table 2). The allowable harvest per day and the hunting season were gradually extended thereafter (Table 1). In fiscal year 1998, 37,800 females and 34,700 males were harvested or culled in eastern Hokkaido (Fig. 5). The doubling of allowable harvest per day and extending the hunting periods for females were very effective in promoting the female harvest. In 2000, the CMPS was revised as an authorized plan by the Environmental Agency of the Japanese Government, and target management areas were extended to central Hokkaido, which consisted of monitoring units 4–7 and the south parts of units 2 and 3; see Fig. 3) along with the expansion of deer distribution and the increase of population size in western and northern Hokkaido (Figs. 2 and 6; Hokkaido Government 2000). At the same time, the Hokkaido Government extended the hunting season to 4 months (November–February) in eastern Hokkaido. The extension of the hunting season to the end of February was very effective in increasing the female harvest in eastern Hokkaido, where 3,658 more females were harvested compared to the previous year. The extension of the
Popul Ecol (2010) 52:373–387
hunting season to the end of February was reversed after 2000, to avoid disturbing the breeding of large birds of prey. The relative population size indices decreased from 123% in 1998 to 79% in 2000, but it did not show further decline thereafter (Fig. 6). Although target management areas were extended to cover all of Hokkaido in 2002, and unlimited female hunting started from 2004 in eastern Hokkaido (Table 1), the female harvest did not increase correspondingly. By 2005, the population size indices had recovered to almost the same level as in 1993 (Fig. 6). Although EPC with aggressive female hunting has been continuing over 10 years since 1998, the population size has still been far beyond the irruption threshold.
Improvement of population estimation for sika deer
381
index to the known number of individuals that have been artificially removed. Thus, hunting is considered as an important large-scale experiment to obtain reliable estimation of population size. Using the feasible sets of parameter values in the resultant simulations for the sika deer population, we re-estimated the 1993 population size to be between 170,000 and 330,000 deer. The lower limit was given by the persistence of males and the upper limit was given by the fact that the population had not continuously decreased in size. In the 2000 CMPS, the Hokkaido Government announced that the 1993 population size had indeed been underestimated. Lesson 3. Population size estimation has large uncertainties. When unpredicted results are obtained, the estimate of the population size should be re-examined first. Hunting can be considered as an important large-scale experiment to obtain reliable estimation of population size.
Harvest-based estimation Harvest-based Bayesian estimation The deer population in eastern Hokkaido for the 1993 fiscal year was first estimated to be between 74,000 and 166,000 individuals (90% CI; Hokkaido Institute of Environmental Sciences 1995). For this estimate, we conducted aerial counts by helicopter at 43 survey units on wintering grounds (Fig. 3; total area was 406.5 km2) between late February and early May in 1993 and 1994, and extrapolated the average density in the survey units into potential wintering grounds in 4 districts of eastern Hokkaido. The harvest exceeded the estimated natural population increase during the first several years under EPC, assuming an annual rate of increase of 16–21% (Kaji et al. 2004). The population size indices were, however, higher than expected, although the population size indices decreased from 1998 to 2000. Moreover, the indices increased again thereafter, and an analysis of population structure, using a population dynamic model, suggested that males should have become extirpated under the culling rates observed in EPC. Because of these contradictions, we suspected that the 1993 population size had been underestimated. The aerial survey might have resulted in under-counting due to the low visibility of deer in dense mixed coniferous– deciduous forests, even though the count was useful in assessing population trends as a population index. Several studies show that ground counts of ungulates also have low precision and often underestimate the population size (e.g., Largo et al. 2008). To solve the inconsistencies among population parameters, Matsuda et al. (2002) proposed a new method to estimate the population size by harvestbased estimation, which requires information about actual trends in relative population size, the rate of natural population increase, and the number of animals harvested (Fig. 5). In this method, the total number of individuals is estimated by examining the response of the population
Using harvest-based estimation, we were able to obtain more reliable estimates of the population size. Uncertainty in the estimation of absolute population size was, however, still large. It is well known that estimates using indices suffer from large observation errors when the probability of observation fluctuates widely; therefore, we applied state– space modeling to the harvest-based estimation combining a generalized linear mixed model (GLMM) with a Bayesian statistical model based on the population dynamic model. We first obtained GLMM estimates of population index by assuming two errors: (1) a lognormal error that was yielded by the ‘local’ random fluctuation of the expected number of observation, and (2) a Poisson error under the given expected number of observations (Yamamura et al. 2008). The GLMM estimates obtained of population index still contain large errors that are caused by the ‘global’ random fluctuation of the probability of observation, that is, the fluctuation of the probability of observation which is synchronous over the whole area. Then, the state–space modeling of harvest-based estimation was applied to the GLMM estimates to remove the influence of such global errors. Bayesian estimation was used for obtaining the maximum likelihood estimates of the state–space model approximately. Then, the dynamics of the index estimated by the state–space modeling (Fig. 6a) became less erratic than that of the index estimated by the simple harvest-based estimation (Fig. 5). This indicates that the state–space modeling that we call the ‘harvestbased Bayesian estimation’ successfully removed the global errors caused by the fluctuation of the probability of observation. Using the model, the 1993 population size ±SE was estimated at 171,000 ± 32,000 in eastern Hokkaido and at
123
382
111,000 ± 63,000 in western Hokkaido. In eastern Hokkaido, the new population estimates showed a decrease from 1998 to 2001, but an increase from around 2002 and eventual recovery to the same population level as the historic peak of 1998 (Fig. 6a). In contrast, the population estimate in western Hokkaido consistently increased even though the number of harvested deer was increasing (Fig. 6b). Lesson 4. Statistical techniques are essential to estimate various variables for populations. A different management purpose may request estimation for a different set of variables or the different level of accuracy for variables to make a management program. It is, therefore, best to customize statistical techniques to each management purpose.
Prevention of lead poisoning in sea eagles A serious side effect of hunting on the sika deer is inadvertent lead poisoning of large birds of prey: Steller’s sea eagles (Haliaeetus pelagicus) and white-tailed sea eagles (H. albicilla). They migrate seasonally between Hokkaido Island and the Kamchatka Peninsula and Sakhalin Island, Russia (Working Group of White-tailed Eagles and Steller’s Sea Eagles 1996; McGrady et al. 2000) and are designated as Japanese natural monuments and listed in the Act on Conservation of Endangered Species of Wild Fauna and Flora in Japan. After lead poisoning in these species was first confirmed in February 1996, incidents increased in association with expanding hunting areas and relaxing hunting regulations. In the 1997 hunting season, a total of 18 sea eagles were found dead (69.2% of total found dead, which were reported by citizens and volunteers, n = 26) and lead bullet fragments were recovered from eagle gizzards in most cases (Uno et al. 2009). It is evident that eagles ingested lead from deer remains in the field (Kurosawa 2000). In 1998, when the CMPS started, the Hokkaido Government established carcass disposal stations and promoted the use of copper bullets. In the 1998 hunting season, however, a total of 26 eagles were found dead from lead poisoning (78.8% of total found dead, n = 33). To prevent further lead poisoning of sea eagles, the Hokkaido Government banned the use of lead bullets in rifles for sika deer hunting from 2000 and in shotguns from 2001, and in rifles for both sika deer and brown bears (Ursus arctos yesoensis) from 2004. Thereafter, the number of sea eagles dying from lead poisoning decreased from 11 (57.9% of total found dead, n = 19) in the winter of 2001/2002 to 2 (8.3% of total found dead, n = 24) in the winter of 2007/ 2008, though one goshawk (Accipiter gentilis) and 2 mountain hawk-eagles (Spizaetus nipalensis) were found dead due to lead poisoning in the winters of 2002/2003 and
123
Popul Ecol (2010) 52:373–387
2003/2004, respectively. In addition, one mountain hawkeagle died from lead poisoning in 2007/2008. Although the Hokkaido Government has banned use of the lead bullets for all hunting since 2004, lead bullets are still legally used and available in other parts of Japan, which may contribute to illegal deer hunting using lead bullets in Hokkaido. Another possibility is the presence of wounded deer, which had already been shot using lead bullets. Lesson 5. The influence of a management program may go beyond the target species (or problem). Unpredicted problems could be invited by the implementation of the program. Accountability for improving the program is essential, when new knowledge is gained and/or an error of past knowledge is found.
Problems with MSY theory for the application to sika deer management Lessons from two sika deer population irruptions: Nakanoshima Island and Cape Shiretoko herds Ungulate populations, when densities are low in favorable habitats or when released from hunting, sometimes increase rapidly to a peak and then crash because of overexploitation of their key food sources (Caughley 1970). Irruptive behavior is common in ungulate populations but is a complex and still poorly documented phenomenon (McCullough 1997). Forsyth and Caley (2006) developed mathematical models to better describe irruptive dynamics of large-herbivore populations and evaluated the dynamics of seven ungulate populations either introduced to new ranges or released from harvesting. A recent study on the population dynamics of the pronghorn (Antilocapra americana) in Yellowstone National Park of the western US also supported the paradigm that irruption is a fundamental pattern in large herbivores with high fecundity and delayed density-dependent effects on recruitments (White et al. 2007). We have observed the irruption processes of two populations of sika deer in detail; an introduced population on Nakanoshima Island (NKI population; Kaji et al. 1988) and a naturally colonized population on Cape Shiretoko (CS population; Kaji et al. 2004; Fig. 8). Both populations built up to peak abundance followed by population crash resulting in significant effects on the vegetation. There were, however, marked differences in post-crash behavior between the two populations. Following the crash in 1984, the NKI herd increased again with a lower growth rate (rm = 0.15 for the period between 1964 and 1984 vs rm = 0.07 for the period between 1986 and 2000; Fig. 8a) and reached a higher peak of the population size in 2001 than the first irruption, while the CS herd showed oscillating
Popul Ecol (2010) 52:373–387 500
383
a Nakanoshima Is.
Number of deer
400
300
200
100
0 1950
1960
1970
1980
1990
2000
2010
were observed as follows; bark-stripping to large elm trees, increasing of bark peeling, elimination of tall plants, formation of browsing line, decreasing of dwarf bamboo, elimination of dwarf bamboo, and invasion of unpalatable plants such as the Aleutian ragwort (Senecio cannabifolius) and an exotic plant, the Bull thistle (Cirsium vulgare) (Kaji et al. 2009). Such a series of vegetation changes associated with high deer density has also been observed widely in eastern Hokkaido (Kaji et al. 2006). Lesson 6. Natural regulation cannot be expected for sika deer where there is a weak density dependence on population growth. Population control is necessary to avoid habitat destruction.
Year 700
Maximum sustainable yields harvest and damages to vegetation
b Cape Shiretoko
600
Number of deer
500 400 300 200 100 0 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
Year Fig. 8 a Population changes of the sika deer population on Nakanoshima Island, Hokkaido, Japan, 1957–2006. Black bar the removal, open circles estimated population sizes, black circles minimum population ground count, and arrows population crashes following peaks (Kaji et al. 1988, updated with unpublished data). b Changes in the sika deer population on Cape Shiretoko, Hokkaido, during 1986– 2005, based on aerial photographic censuses (Kaji et al. 2004, updated with unpublished data)
behavior without decline in the peak abundance (1986 vs 2003; Fig. 8b). The intrinsic rates of natural increase (rm) of the two populations ranged between 0.15 and 0.19 (doubling time is about 4 years). Because the two populations did not show evident density effects on population growth rate just before the populations crashed, natural regulation cannot be expected to prevent damage to natural vegetation. Although density-dependent resource limitation through harsh conditions in winter was the common limiting factor in peak densities for both populations, the carrying capacity differences and resource gaps between summer and winter might generate the difference in fluctuation patterns between the two populations (Kaji et al. 2009). On Nakanoshima Island and Cape Shiretoko, associated with population irruption, a series of vegetation changes
Density dependence has been an important theoretical base for deer management, because the MSY model that harvest is compensated by recruits assumes a feedback mechanism based on density dependence (White and Bartmann 1997). The idea of managing a deer population at the density that provides the MSY has been a prevailing influence in the management of European and North American populations, particularly where management goals stress production for hunting. In other parts of the USA and Europe there may be less tolerance to the MSY approach now because of concerns about overabundance in deer (McShea et al. 1997; Linnell et al. 1998; Coˆte´ et al. 2004). Ecological carrying capacity, defined as the maximum population that an environment can support without detrimental effects (Caughley and Sinclair 1994), is one of the equilibrium points as represented by K of the logistic equation. MSY is the maximum number of deer that can be harvested from the population on a continual basis. The population level that produces the MSY is 50% of K and has been referred to as Economic Carrying Capacity (Caughley 1979), or I-Carrying Capacity (Macnab 1985), where I-Carrying Capacity equates to the inflection point on the sigmoid curve and is the density where the number of recruits is maximum. McCullough (1984) stated that desirable goals between MSY and K are inherently self-correcting. If overharvest or a severe winter reduces a population below a goal, the smaller population would produce a larger number of recruits and therefore would quickly return to the goal. Conversely, goals below MSY may make a population unstable because the population size is likely to suffer higher demographic stochasticity owing to small population size (McCullough 1984). Since MSY (50% of K) is far below ecological carrying capacity (K), MSY would at first seem to satisfy the conditions that prevent negative impacts on vegetation.
123
384
Cervid species have expanded their range and increased dramatically in abundance particularly in the developed world in recent decades, and they have caused not only major economic losses in agriculture and forestry but also serious negative impact on ecosystems (McShea et al. 1997; Coˆte´ et al. 2004). This might be caused by any of the following reasons; (1) deer populations are close to K due to insufficient population management, (2) K/2 may be too high to prevent damage (deCalesta and Stout 1997), and/or (3) MSY theory lacks reality because forest fragmentation and croplands in the developed world may have increased the landscape’s ability to support deer, as explained by the ‘‘fragmentation–nutrition hypothesis’’ (Sinclair 1997). The problem with MSY theory is that it assumes density dependence. We are not able to determine K if a population shows no significant density dependence. Even if a population showed density dependence, if it is weak, density effects on population parameters would be observed only when the population increased to a high density that is close to K. Measuring K has, therefore, proven extremely difficult, because we usually do not want the population to ever reach K (Mysterud 2006). It is known that sika deer do not exhibit clear density-dependent effects at relatively low densities (Putman and Clifton-Bligh 1997). Using a sika deer population reconstructed by cohort analysis, Ueno et al. (2010) demonstrated a density-dependent decrease in recruitment through its negative effect on fawn survival. The density effect was, however, not strong enough to achieve population regulation. Ungulate species in North America occupying high productivity areas with relatively stable climates (e.g., the white tailed deer, Odocoileus virginianus) tend to conform to the logistic model, whereas the species living in moderately productive environments (e.g., the mule deer, O. hemionus and the elk, C. elaphus), or living in low productivity environments (e.g., the bighorn sheep, Ovis canadensis nelson and the desert mule deer, O. h. crooki) show continuous population growth at relatively constant rates and most density-dependent responses become detectable only when N grows to be as high as K (McCullough 1999). Thus, the MSY point is located above 50% of K in many ungulate species of which population growth rate shows a convex curve against their density. We demonstrate, in the two case studies on population irruption of sika deer, that negative impacts on vegetation occur before density dependence becomes observable. In other words, habitat destruction may have progressed up to an undesired level by the time we get enough demographic data to measure K. To prevent habitat destruction, therefore, we have to set a target density for management without information about K. Relative deer density (RDD, based on deer density relative to K) was introduced as a concept to integrate white-
123
Popul Ecol (2010) 52:373–387
tailed deer management with ecosystem management (deCalesta and Stout 1997). We evaluated RDD (DD/ K) 9 100 and its effect on natural vegetation on Nakanoshima Island and Cape Shiretoko. K on NKI was estimated based on the relationship between the rate of population change [(Nt?1 - Nt)/Nt?1] and estimated population size during the winter 1980–1984 in the form Y ¼ 0:0021X þ 0:6381 R2 ¼ 0:9221; P\0:05 ; where Y is the rate of population change as a function of population size X. The rate of population change decreased as population size increased. We obtained K = 304 deer (61/km2) when the rate of population change was 0. NKI population reached the first peak of 299 deer (57.5/km2) in the autumn 1983, which was close to K, and crashed in the following winter. The CS herd irrupted and reached peaks three times during 1986–2005. These estimated peak population sizes were 592 deer in 1998, 626 in 2003, and 603 in 2005. Thus, we estimated averaging K is 121/km2 (range 118–125/km2) on CS. On NKI, the pasture at the central part of the island consisted of forbs, grasses and tall plants over 100 cm in height, and there were no obvious effect of grazing on the vegetation in 1980 when deer density was 31.5/km2 (54%K) in 1980 (Kaji et al. 1988). On CS, woody vegetation showed signs of overbrowsing when deer density was [15/km2 (13%K) in 1987, and considerable overbrowsing of dwarf bamboo and woody plants occurred when estimated deer density was 62/km2 (51%K) in 1995 (Kaji et al. 2004). Both populations showed that signs of overbrowsing appeared when the deer population over 50%K (Table 3). Thus, the two irruptive populations of sika deer (NKI and CS herds) show that the density level of MSY management (50%K) may allow deer to have significant negative impacts on natural vegetation. Much lower deer density (e.g., \25% of K) is needed to maintain viable populations of certain sensitive plant species (Alverson et al. 1988; deCalesta and Stout 1997; Sinclair 1997) or to keep deer populations within a sociological carrying capacity—a socially tolerable density level (Sinclair 1997). Our lower limit (%P-) is set based on a critical population size, which should be larger than the vulnerable size (\1,000) in criterion D1 of red list categories (IUCN 1994). Threshold harvesting strategies to avoid damage to vegetation From the above considerations, it is clear that MSY theory is incompatible with many management goals and K/2 is too high for ecosystem management of cervid populations with weak density dependence. Instead of the MSY approach, we developed a threshold harvesting approach
Popul Ecol (2010) 52:373–387
385
Table 3 Relationship between relative deer density (RDD) and its effects on vegetation on Nakanoshima Island and Cape Shiretoko, Hokkaido, Japan RDD
Nakanoshima Island
Cape Shiretoko
Low \20%K
No data
Impacts are minor to moderate: small trees disappeared and browsing line appeared
Low-moderate 20–39%K
No data
No data
Moderate-high 40–59%K
Impacts are obvious: saplings disappeared, dwarf bamboos and small trees declined on wintering grounds
Impacts are obvious: dwarf bamboo and tall plants disappeared, unpalatable plants increased
High C60%K
Great impacts: tall grass disappeared, browsing line appeared, dwarf bamboos disappeared, large trees were barked, and unpalatable plants increased
Great impacts: large oak trees were barked
Classification of RDD (low, moderate, high) based on deCalesta and Stout (1997)
(Lande et al. 1995, 1997) as feedback management; when the number above some threshold population is harvested while below the threshold there is no harvest. Feedback management for the sika deer population set three population thresholds and four options of different hunting pressures which are adopted according to relative population size (Table 2; Fig. 7). Associated with the increase of relative population size (Fig. 7), in addition to agriculture and forestry, the damage of natural forests became significant in various places of Hokkaido. Road and railway kill of deer have also increased. On the analogy of the relationship between population abundance and habitat in NK and CP, these conditions indicate that population density levels in eastern Hokkaido have nearly equaled or exceeded 50%K. Thus, our upper limit (%P?) and our target level (%P*) are considered as well below the MSY level (K/2). Another important point that should be considered for ecosystem management of sika deer in Hokkaido is the large gap between summer and winter range carrying capacity as previously described. Sika deer in eastern Hokkaido have a large-scale seasonal migration and large numbers of deer concentrated in limited wintering ranges, where significant habitat deterioration occurs. Since density-dependent resource limitation with winter weather conditions determine the population size of sika deer in Hokkaido (Kaji et al. 2009), the management goal should be set based on the winter range condition. Lesson 7. MSY theory is unrealistic for cervid populations with weak density dependence because of the difficulty of measuring K, and the fact that the density level set by MSY may be too high to prevent damage to agriculture and/or the ecosystem. More intensive management of a deer population at lower densities is required for ecosystem management based on different threshold harvesting strategies.
Conclusion: hunting as resource management and culling for ecosystem management The greatest change in deer management during the last decade is the one from sport hunting to damage control. The Hokkaido Government has been promoting aggressive female culling through EPC to reduce damage caused by high density deer populations. The program once succeeded in reducing the size of the deer population and amount of agriculture/forestry damage incurred between 1998 and 2000 in eastern Hokkaido, but deer abundance is still high (about 200,000 deer) due to reductions in the harvest, despite the relaxation of hunting regulations. In 2008, the relative population index in eastern Hokkaido is 130 (about 260,000 deer), which is roughly equivalent to the value in 1998 when the population reached a peak. Thus, at least 38,000 female deer, which is the actual result in 1998, should be harvested to reduce the population again. We could not expect this implementation problem (the difficulty of harvest) when we originally designed the CMPS plan in 1998. In 2000, a total of 10,000 registered hunters in Hokkaido (6% of hunters in Japan) harvested 70,000 deer ([50% of the total harvest); each hunter harvested an average of 7 deer. Although the hunting pressure on deer in Hokkaido is the highest in Japan, sika deer populations have expanded to cover nearly all their potential range, and occur in natural reserves, including alpine meadows, where natural vegetation is negatively affected. In national parks and nature reserves, where conservation of the natural environment is the highest priority, overabundance of deer is a pressing and major concern with regard to conserving vulnerable rare plant species. Another concern is that, while much deer management has been focused on damage control in rural areas, deer have recently been invading urban areas. What we are trying to achieve now is landscape scale
123
386
management (at the scale of the deer populations) which can be part of an ecosystems-based approach to natural resources management. At the same time, it is clear that different management challenges may be faced by different elements within a landscape (for example, a nature reserve as opposed to an agricultural or peri-urban area). Since the hunter population is decreasing, encouragement for utilization of sika deer as a natural resource, and hunter training and support, are urgently required to maintain hunting systems. A new community-based management trial, combining recreational hunting and hunter education for rural areas, has been underway in Nishiokoppe Village, Hokkaido, from 2004 (Igota and Suzuki 2008). In the CMPS (third period: 2008–2011), to encourage harvesting, sika deer are regarded as a valuable natural resource as venison, and sustainable resource management must be considered: (http://www.pref.hokkaido.lg.jp/NR/ rdonlyres/1B7408C3-2E60-43E4-9423-3079F8517AE6/0/ 3rd_plan.pdf.). In managing deer as a resource, hunters may be required to have relatively high levels of hunting efficiency or the resulting densities would be too high for ecosystem management. To avoid negative impacts on natural vegetation in national parks, and to conserve biodiversity, much lower densities of deer should be set for target areas. Since hunting efficiency would be too low for volunteer hunters in those areas, it is important to support the development of professional hunters for culling operations. Hunting as resource management and culling for ecosystem management should be synergistically combined under adaptive management. Acknowledgments We are very grateful to T. Akasaka, T. Fujimoto, K. Tamada, T. Kurumada and H. Hirakawa for supporting our work and stimulating discussion, while over the years, N. Ohtaishi and D. R. McCullough have consistently encouraged us to develop sika deer management in Hokkaido. We also thank Ed Dyson, the editor and anonymous referees who gave us valuable comments and corrected English on the manuscript, and T. Suzuki for drawing figures. This work was supported in part by a Grant-in-Aid for Scientific Research (A) No. 21248019 from the Ministry of Education, Culture, Sports, Science and Technology to K.K.
References Alverson WS, Waller DM, Solheim SL (1988) Forest too deer: edge effects in northern Wisconsin. Conserv Biol 2:348–358 Caughley G (1970) Eruption of ungulate populations with emphasis on Himalayan thar in New Zealand. Ecology 51:53–72 Caughley G (1979) What is this thing called carrying capacity? In: Boyce MS, Hayden-Wing LD (eds) North American elk: ecology, behavior, and management. University of Wyoming Press, Laramie, pp 2–8 Caughley G, Sinclair ARE (1994) Wildlife ecology and management. Blackwell, Oxford Christensen NL, Bartuska AM, Brown JH, Carpenter S, Antonio CD, Francis R, Franklin J, MacMahon JA, Noss RF, Parsons DJ,
123
Popul Ecol (2010) 52:373–387 Peterson CH, Turner MG, Woodmansee RG (1996) The report of the ecological society of America committee on the scientific basis for ecosystem management. Ecol Appl 6:665–691 Coˆte´ SD, Rooney TP, Tremblay JP, Dussault C, Waller M (2004) Ecological impact of deer overabundance. Annu Rev Ecol Evol Syst 35:113–147 deCalesta DS, Stout SL (1997) Relative deer density and sustainability: a conceptual framework for integrating deer management with ecosystem management. Wildl Soc Bull 25:252–258 Forsyth DM, Carey P (2006) Testing the irruptive paradigm of largeherbivore dynamics. Ecology 87:297–303 Government Hokkaido (1998) Conservation and management plan for sika deer (Cervus nippon) in eastern Hokkaido. Hokkaido Government, Department of Health and Environment, Sapporo, Japan (in Japanese) Hokkaido Government (2000) Conservation and management plan for sika deer (Cervus nippon) in Hokkaido. Hokkaido Government, Department of Health and Environment, Sapporo, Japan (in Japanese) Hokkaido Institute of Environmental Sciences (1994) Distribution of sika deer and brown bear on Hokkaido. Hokkaido Institute of Environmental Sciences, Sapporo (in Japanese) Hokkaido Institute of Environmental Sciences (1995) Results of a survey related to sika deer and brown bear in Hokkaido. Hokkaido Institute of Environmental Sciences, Sapporo (in Japanese) Hokkaido Institute of Environmental Sciences (1997) Results of a survey related to sika deer and brown bear in Hokkaido. Hokkaido Institute of Environmental Sciences, Sapporo (in Japanese) Igota H, Suzuki M (2008) Community-based wildlife management: a case study of sika deer in Japan. Human Dimens Wildl 13:416– 428 Igota H, Sakuragi M, Uno H, Kaji K, Kaneko M, Akamatsu R, Maekawa K (2004) Seasonal migration patterns of female sika deer in eastern Hokkaido, Japan. Ecol Res 19:169–178 Inukai T (1933) Review on extirpation of wolves in Hokkaido. Shokubutu to Dobutsu (Plants and animals) 1:1091–1098 (in Japanese) Inukai T (1952) The sika deer in Hokkaido and its raise and decline. Hoppo Bunka Kenkyu Hokoku (The Report of Northern Cultural Research) 7:1–45 (in Japanese) IUCN (World Conservation Union) (1994) IUCN red list categories. IUCN, Gland Kaji K, Koizumi T, Ohtaishi N (1988) Effects of resource limitation on the physical and reproductive condition of sika deer on Nakanoshima Island, Hokkaido. Acta Theriol 33:187–208 Kaji K, Miyaki M, Saitoh T, Ono S, Kaneko M (2000) Spatial distribution of an expanding population on Hokkaido Island, Japan. Wildl Soc Bull 28:699–707 Kaji K, Okada H, Yamanaka M, Matsuda H, Yabe T (2004) Irruption of a colonizing sika deer population. J Wildl Manage 68:889– 899 Kaji K, Miyaki M, Uno H (eds) (2006) Conservation and management of sika deer in Hokkaido. Hokkaido University Press, Sapporo (in Japanese) Kaji K, Takahashi H, Okada H, Kohira M, Yamanaka M (2009) Irruptive behavior of sika deer. In: McCullough DR, Takatsuki S, Kaji K (eds) Sika deer: biology and management of native and introduced populations. Springer, Tokyo, pp 421–436 Kaneko M, Kaji K, Ono S (1998) An analysis of the change of distribution accompanying the change of habitat of Hokkaido sika deer. Honyurui Kagaku (Mammalian Science) 38:49–59 (in Japanese) Kurosawa N (2000) Lead poisoning in Steller’s Sea Eagles and White-tailed Sea Eagles. In: Ueta M, McGrady MJ (eds) First
Popul Ecol (2010) 52:373–387 symposium on Steller’s and white-tailed sea eagles in east Asia. Wild Bird Society of Japan, Tokyo, pp 107–109 Lande R, Engen S, Sæther B-E (1995) Optimal harvesting of fluctuating populations with a risk of extinction. Am Nat 145:728–745 Lande R, Sæther B-E, Engen S (1997) Threshold harvesting for sustainability of fluctuating resources. Ecology 78:1341–1350 Largo E, Gaillard J-M, Festa-Bianchet M, Toı¨go C, Bassano B, Cortot H, Farny G, Lequette B, Gauthier D, Martinot J-P (2008) Can ground counts reliably monitor ibex Capra ibex populations? Wildl Biol 14:489–499 Linnell JDC, Duncan P, Andersen R (1998) The European roe deer: a portrait of a successful species. In: Andersen R, Duncan P, Linnell JDC (eds) The European roe deer: the biology of success. Scandinavian University Press, Oslo, pp 11–22 Lotka AJ (1925) Elements of physical biology. (Reprinted by Dover in 1956 as Elements of mathematical biology) Macnab J (1985) Carrying capacity and related slippery shibboleths. Wildl Soc Bull 13:403–410 Matsuda H, Kaji K, Uno H, Hirakawa H, Saitoh T (1999) A management policy for sika deer based on sex-specific hunting. Res Popul Ecol 41:139–149 Matsuda H, Uno H, Tamada K, Kaji K, Saitoh T, Hirakawa H, Kurumada T, Fujimoto T (2002) Harvest-based estimation of population size for sika deer on Hokkaido Island, Japan. Wildl Soc Bull 30:1160–1171 McCullough DR (1984) Lessons from the George Reserve, Michigan. In: Halls LK (ed) White-tailed deer ecology and management. Wildl Manage Institute, Stackpole Books, Harrisburg, pp 211– 242 McCullough DR (1997) Irruptive behavior in ungulates. In: McShea WJ, Underwood HB, Rappole JH (eds) The science of overabundance: deer ecology and population management. Smithsonian Institution Press, Washington DC, pp 69–98 McCullough DR (1999) Density dependence and life-history strategies of ungulates. J Mamm 80:1130–1146 McGrady MJ, Ueta M, Potapov E, Utekhina I, Masterov VB, Fuller M, Seegar WS, Ladyguin A, Lobkov EG, Zykov VB (2000) Migration and wintering of juvenile and immature Steller’s sea eagles. In: Ueta M, McGrady MJ (eds) First symposium on Steller’s and white-tailed sea eagles in east Asia. Wild Bird Society of Japan, Tokyo, pp 83–90 McShea WJ, Underwood HB, Rappole JH (eds) (1997) The science of overabundance: deer ecology and population management. Smithsonian Institution Press, Washington DC Morellet N, Gaillard JM, Hewison AJM, Ballon P, Boscardin YVES, Duncan P, Klein F, Maillard D (2007) Indicators of ecological change: new tools for managing populations of large herbivores. J Appl Ecol 44:634–643 Mysterud A (2006) The concept of overgrazing and its role in management of large herbivores. Wildl Biol 12:129–141 Nagata J, Masuda R, Kaji K, Kaneko M, Yoshida MC (1998) Genetic variation and population structure of the Japanese sika deer (Cervus nippon) in Hokkaido Island, based on mitochondrial D-loop sequences. Mol Ecol 7:871–877 Ohtaishi N, Kaji K, Mano T (eds) (1990) Proceeding of deer and bear forum Hokkaido 1990. Wildlife Information Center, Sapporo (in Japanese)
387 Porter WF, Underwood HB (1999) Of elephants and blind men: deer management in the US National Parks. Ecol Appl 9:3–9 Putman RJ, Clifton-Bligh JR (1997) Age-related body weight, fecundity and population change in a south Dorset sika population (Cervus nippon); 1985–1993. J Nat Hist 31:649–660 Sakuragi M, Igota H, Uno H, Kaji K, Kaneko M, Akamatsu R, Maekawa K (2003) Seasonal habitat selection of an expanding sika deer Cervus nippon population in eastern Hokkaido, Japan. Wildl Biol 9:141–153 Sinclair ARE (1997) Carrying capacity and the overabundance of deer: a framework for management. In: McShea WJ, Underwood HB, Rappole JH (eds) The science of overabundance: deer ecology and population management. Smithsonian Institution Press, Washington DC, pp 380–394 Tanaka S (1982) The management of a stock-fishery system by manipulating the catch quota based on the difference between present and target stock level. Bull Jpn Soc Sci Fish 48:1725– 1729 Tawara H (1979) A review on the history of nature preservation in Hokkaido Island. Hokkaido University Press, Sapporo (in Japanese) Thompson WL, White GC, Gowan C (1998) Monitoring vertebrate populations. Academic, New York Ueno M, Kaji K, Saitoh T (2010) Culling versus density effects in the management of a deer population. J Wildl Manage (in press) Uno H, Kaji K (2000) Seasonal movements of female sika deer in eastern Hokkaido, Japan. Mamm Study 25:45–57 Uno H, Kaji K, Saitoh T, Matsuda H, Hirakawa H, Yamamura K, Tamada K (2006) Evaluation of relative density indices for sika deer in eastern Hokkaido, Japan. Ecol Res 21:624–632 Uno H, Kaji K, Tamada K (2009) Sika deer irruptions and their management on Hokkaido Island, Japan. In: McCullough DR, Takatsuki S, Kaji K (eds) Sika deer: biology and management of native and introduced populations. Springer, Tokyo, pp 405–419 Volterra V (1926) Fluctuations in the abundance of a species considered mathematically. Nature 118:558–560 Walters CJ (1986) Adaptive management of renewable resources. McMillan, New York White GC (2001) Why take calculus? Rigor in wildlife management. Wildl Soc Bull 29:380–386 White GC, Bartmann RM (1997) Density dependence in deer populations. In: McShea WL, Underwood HB, Rappole JH (eds) The Science of overabundance. Smithsonian Institution Press, Washington DC, pp 120–135 White PJ, Bruggeman JE, Garrott RA (2007) Irruptive population dynamics in Yellowstone pronghorn. Ecol Appl 17:1598–1606 Working Group of White-tailed Eagles and Steller’s Sea Eagles (1996) Wintering status of Steller’s sea eagles and white-tailed eagles in northern Japan. In: Survey of status and habitat conditions of threatened species. Environment Agency, Tokyo, Japan, pp 1–9 (in Japanese) Yamamura K, Matsuda H, Yokomizo H, Kaji K, Uno H, Tamada K, Kurumada T, Saitoh T, Hirakawa H (2008) Harvest-based Bayesian estimation of sika deer populations using a state-space model. Popul Ecol 50:131–144
123
