In uneven-aged mixed conifer–broadleaf forests in northern Japan, monarch birch (Betula maximowicziana
Regel), castor aralia (Kalopanax septemlobus
(Thunb.) Koidz), and Japanese oak (Quercus crispula
Blume) are important producers of high commercial value timber, which is used in the veneer and furniture industries. The supply of high-quality timber from these tree species is exclusively dependent on the cutting of large trees within the mixed forests. Large-sized trees in mixed forests contribute to the structural heterogeneity, dynamics, and functions of the forest ecosystem [1
], a large fraction of aboveground biomass and carbon storage [3
], and play important roles in the rate and pattern of natural regeneration [4
]. In addition, high-value timber species may be subjected to excessive and illegal logging [5
], and the number of large-sized high-value trees is declining [7
] in many regions of the world. Understanding the sustainability of these species will facilitate forest management, conservation, and the simulation of various silvicultural practices.
The achievement of sustainability from the use of various forest management practices is a central precept of forestry and is therefore central to all silvicultural systems [8
]. Uneven-aged forest management or selection system has gained growing interest in many parts of the world due to its stability in forest stand structures [9
], and there has been increasing criticism for even-aged forestry, wherein the whole forest area is clear-cut and regenerated artificially. Furthermore, sustainable forest management (SFM) has been encouraged as a guiding principle in forest management [13
] and uneven-aged forest management is assumed to achieve greater sustainability in forest resource management in comparison with even-aged forest management [8
]. It is sometimes referred to as close-to-nature forest management [15
], which implies the achievement of a form of silviculture that emulates natural processes resulting in stand structures that are natural, and it promotes natural processes such as soil productivity maintenance, nutrient cycling, and biodiversity [16
In uneven-aged mixed conifer–broadleaf forests in northern Japan, the selection system has been practiced as a common management system since the early twentieth century [17
]. In fact, the selection management system attempts to mimic natural disturbances through the use of various management practices. General expectations of the use of selection management are the increased growth, recruitment, and survival of remaining trees [18
]. Several research attempts have been made in different parts of the world to investigate the impacts of selection management on the remaining forest stand demographic parameters, i.e., growth, recruitment, and mortality (e.g., [10
]), as well as species composition and stand structure (e.g., [14
]). Many previous studies in mixed conifer–broadleaf forests in northern Japan have also assessed the impact of selection management on the growth, recruitment, and mortality of the remaining forest stand (e.g., [18
]). However, few studies have examined the sustainability of forest stands, especially high-value timber species. Information on the sustainability of high-value timber species within mixed forests is important for forestry practitioners, especially when the goal of forest management is to manage certain species. Moreover, understanding the sustainability of uneven-aged mixed forests is useful for forest management decision [29
] because it helps to determine whether or not a specific stand structure should be maintained [30
The assessment of sustainability in uneven-aged mixed forest is relatively difficult because forests are slow growing and it may take several decades to examine the long-term impacts of any given forest management activities. It requires criteria and indicators that can be measured during stand development [8
] after forest management activities. The availability of long term forest measurement data, therefore, is an important source of information [32
], not only for providing information of forest stand dynamics but also to assess the sustainability of forest stands subjected to various natural and anthropogenic disturbances.
The parameters derived from long-term forest measurement data would be useful for assessing the sustainability of a forest management system. O’Hara et al. [8
] compared forest stand parameters such as stocking (tree density and basal area), species diversity, stand structure, and increment between even-aged forest and a selection system in Central Europe as measures of sustainability using over 90 years of forest measurement data. In addition, Schuler [21
] examined the species composition, diversity, and growth of tree species in mixed mesophytic forest in the USA after 50 years of partial harvesting. In mixed conifer–broadleaf forest in northern Japan, Yoshida et al. [18
] assessed the dynamics of a forest stand after 20 years of selection harvest. However, these studies examined long term changes in the stocking and demographic parameters of major tree species or stand level stocking, species diversity, and stand structure. The sustainability of high-value timber species after selection harvest has not been widely studied. Understanding the sustainability of high-value timber species will be useful for the reliable application of a selection system and species-specific forest management or a single-tree management system, which was recommended in previous studies (e.g., [7
The aim of this study is, therefore, to assess the sustainability of high-value timber species in mixed forests managed under selection systems. Using 48 years of measurement data, we derived the stocking, demographic parameters, and species proportion of high-value timber species as measures of sustainability. To reach the objective, firstly, we assessed the changes in stocking and demographic characteristics of high-value timber species over time. Secondly, the changes in the species proportion of high-value timber species were assessed. In addition, we also showed how the forest stand structure was changing by assessing the sustainability measures of all conifer and broadleaf species.
In this study, we investigated the long-term changes in stocking, demographic characteristics, and species proportion of high-value timber species as measures of sustainability. We used these parameters as criteria and indicators to evaluate the sustainability of high-value timber species as proposed by O’Hara et al. [8
]. However, they used these criteria and indicators to evaluate the stand level sustainability of even-aged and uneven-aged forest management. The changes or consistency of these parameters would be useful for establishing sustainable forest management by adjusting the tree marking for harvesting.
The main common characteristics of the targeted high-value timber species in this study were increases in N and BA. The increasing trend was also observed in the total N and total BA of forest stands. An increase in the N of forest stands managed under the selection system has occurred in the last few decades in other parts of the world with different environment and forest types. For example, Klopcic et al. [20
] found an increased total number of trees in their study in Slovenia. Compared with an unmanaged stand, a higher mean density and basal area in the managed stand was also reported by Young et al. [23
] in USA. Moreover, a decreasing tree density and basal area in an unmanaged stand was reported by Ediriweera et al. [24
] in their study in mixed-dipterocarp forests over a 40-year period.
One possible reason for increasing the N and BA of forest, including those of high-value timber species, would be due to some major disturbance, including natural (e.g., strong typhoon) and anthropogenic (e.g., selection harvest) factors [26
]. In the mixed conifer–broadleaf forest in northern Japan, a large typhoon occurred, causing widespread canopy opening [40
] in some plots. Selection harvesting was carried out three to four times during the 48 years period. The increasing trend of both total N and BA of forest stand was greatly contributed by broadleaf species. These increasing trends in both N and BA are consistent with previous studies in mixed conifer–broadleaf forests in northern Japan such as that done by Yoshida et al. [18
]. Those studies also highlighted the increasing trend of broadleaf tree density in mixed conifer–broadleaf forests managed under selection system.
Reversed J-shaped diameter distributions were observed for all target high-value timber species (Figure 2
). Diameter distribution curves indicated an increasing number of smaller diameter class trees for all species across census periods. Owari et al. [35
] also reported a large number of smaller diameter class trees in northern Japanese mixed conifer–broadleaf forests. More fluctuation in size structure of monarch birch and Japanese oak were observed.
Similar to N and BA, the mean BAI of high-value timber increased over time. The increasing trend was also observed for the mean stand BAI even though it was not significantly different in the first three census periods. The results of regression modeling suggested temporal trends of BAI for high-value timber species and the total BAI of forest stands. In terms of the stand level BAI, a significant positive trend was observed. In northern mixed conifer–broadleaf forests, similar results have been reported for the broadleaf species (e.g., [18
The mean N-mor across census periods showed no significant difference for all high-value timber species. However, Japanese oak N-mor increased over time. Furthermore, no significant increasing or decreasing trends in N-rec were observed for high-value timber species across census periods in this study. These results are also consistent with previous studies by Hiura et al. [41
]. For all target species, a significantly lower N-rec was observed in the first and last census periods, while a higher N-rec was observed in the second, third and fourth census periods. This trend can also be observed for the total N-rec of all species. This pattern may also be explained by a large disturbance caused by a large typhoon in 1981 (second census period) which created a canopy opening causing high light availability [42
] in some plots, favoring natural regeneration. As a result, higher N-rec in these plots would be expected following a large natural disturbance. This was also highlighted by previous studies in the mixed conifer–broadleaf forest in Northern Japan. Yoshida et al. [18
], for example, reported that a higher N-rec of castor aralia would be expected after a disturbance, and harvesting treatment in their study. In addition, Takahashi et al. [43
] reported that Japanese oak may tend to regenerate after a large disturbance before the establishment of other species.
The results in this study also clearly show another important temporal trend of the forest stand. The forest composition and structure in the study area changed over time after the first census period (1968 to 1978). The proportion of conifer in the first census period was 48.35%, and it decreased to 33% in the last census period. An increasing broadleaf proportion might contribute to an increased N of high-value timber species (Table 2
). Broadleaf N and BA exceeded those of conifer after the first census period, wherein the N and BA of conifer were larger than those of broadleaf. Even though there might be several reasons for this, one of the possible reasons for decreased conifer N and BA would be due to a bias in tree marking for harvesting. Table 6
showed that the total harvested basal area for conifer always exceeds that of broadleaf in all census periods. Owari et al. [36
] indicated that even though spatially unbiased tree marking was obtained, tree species that were marked for selection harvest in our study area was mostly conifer species. In addition, they observed that the trees marked for harvesting were larger than the unmarked trees.
In terms of BAI, conifer showed a decreasing trend over time (Table 3
). Even though the mean BAI of conifer exceeded the mean BAI of broadleaf throughout the census periods, both were almost similar in the last census period (0.379 m2
/ha for conifer and 0.369 m2
/ha for broadleaf). In addition, the mean BAI of conifer was significantly lower in the last census period than in the first four periods. Yoshida et al. [18
] reported that the growth of shade-intolerant broadleaf species was less than that of most common conifer species (i.e., A. sachalinensis
), even after selection harvesting. However, our study found that even though a larger mean BAI was observed for all conifer species, its BAI showed a decreasing trend while the total BAI of broadleaf species exhibited increasing trend. Harvesting of more conifer trees may contribute lower mean value in total BAI. Our results are in line with a previous study by Hiura et al. [41
] in mixed conifer–broadleaf forest in northern Japan. According to this trend, the future forest composition of the study area would be more of a broadleaf-dominating type in terms of both N and BA. A similar decreasing trend in the conifer proportion has been reported in different regions [20
The results of the sustainability measures in this study revealed that there have been inconstancies in these measures over time. However, such inconsistencies could relate to a number of reasons. A recent study [41
] in mixed conifer–broadleaf forest with very little human disturbance in northern Japan revealed that changing climate conditions such as an increased temperature, precipitation, and decreased snowfall and snow cover period have led a to reduction in growth rate of conifer and an increasing in that of broadleaf species. Moreover, these inconsistencies in sustainability measures due to a changing climate have been widely reported in different regions [44
]. Similar to the results by O’Hara et al. [8
], the results of this study indicate that a single-tree selection system is more of a dynamic entity.
The sustainability measures described in this study would be useful for adjusting forest management activities, and various silvicultural activities, which could lead to consistency in sustainability measures in different forest types. Through the understanding of sustainability measures used in this study, forest management can maintain the stocking of uneven-aged forest stand over time, BAI can be balanced by tree removals, and recruitment can be assessed whether it is sufficient. In addition, it would provide information for forest management operations such as stocking control, which is central to uneven-aged silviculture. Many stocking control approaches have been developed including reversed J-shape diameter distribution, selection system or plenter system, stand density index, and leaf area allocation, etc. [46
]. Sustainability measures can be achieved through these stocking control approaches by removing those trees that surpass or maintain those of limited numbers in a certain diameter class.