Forests provide various outcomes and utilities such as carbon storage, water production, soil protection, hunting, amenities and recreational facilities for the society besides conventional wood and non-wood forest products. Due to a variety of expectations by the society from the forest resources over the last decades, multiple-use forest management has become the fundamental component of national forest policy. Additionally, managing forest ecosystems that provide multiple goods and services on a sustainable basis is of great challenge and can be achieved through holistic management of ecological, economical and socio-cultural values of forest ecosystems (Christensen et al., 1996; Davis et al., 2001; Baskent et al., 2008a) In this way, the integration of water production and carbon sequestration into forest management planning becomes a significant challenge of contemporary research endeavors.

Forest ecosystems cover wide range of areas on earth and as such they are one of the most important natural assets affecting the management of water resources. Therefore, forest ecosystems have significant function in the management of water resources. However, water resources in terms of quantity and quality are affected by anthropogenic activities as well as natural disturbances. Currently, the perceivable effects of climate change have raised great attention to disclose the water-forest relationship in forest ecosystems under multiple-use forest management concept (Birot et al., 2011). Some studies figured out that change in forest ecosystems such as reforestation, afforestation and deforestation can affect water quantity and quality. For example, researchers such as Dijk & Keenan (2007) found out that increasing plantation age affects the amount of water negatively in afforested watersheds, raising attention to evaluate effects of various rotation ages for multiple-use in forestry. By using linear regression analysis, Sahin & Hall (1996) showed that a 10% reduction in cover caused 20-25 mm and 17-19 mm increase in water yield from conifer and deciduous forest, respectively. Thus, minimum harvesting age particularly in managed forest has become an effective indicator for the sustainable management of forest ecosystems for multiple values.

Over the last decade, the role of forest ecosystems in carbon sequestration has been well recognized because of their critical importance in controlling the global carbon cycle. Fossil-fuel burning and loss of terrestrial vegetation, for instance, have caused the atmospheric carbon dioxide to increase (Huston & Marland, 2003). The carbon dynamics in forests are largely influenced by periodic disturbances of wildfire, insects, disease, degradation, harvesting and over-harvesting (Kurz et al., 2008; Brown, 2002). Besides, increasing gas concentration created greenhouse effects in the atmosphere that has become one of the factors on global climate change. In this respect, Kyoto protocol prepared by the UN Framework Convention on Climate Change (UNFCCC) is a critical instrument, highlighting the importance of forest ecosystems for meeting the limits for carbon emissions (Brown et al., 1999). This protocol aims to minimize net greenhouse gas emissions through maximizing afforestation-reforestation activities and minimizing deforestation of forest ecosystems (Huston & Marland, 2003). Several management practices such as intensive silviculture interventions and various rotation ages can be effective on the amount of stored carbon in forest ecosystem (Backéus et al., 2005; Swanson, 2009). Moreover, sustainable land use and forest ecosystem planning are two key factors on decreasing the level of greenhouse gases in the atmosphere (Hu & Wang, 2008).

In recent years, forest management planning has progressed from classical planning approach to ecosystem or landscape management approach that harmonizes various potential conflicts between goods and services. The new approach accommodates the sustainable management of multiple forest values. However, when various forest values were integrated into forest management plans, it is quite difficult to estimate what minimum harvesting ages would be optimal for the holistic management of the resources. Determining optimal minimum harvesting ages for sustainable utilization of multiple values from forests is a great challenge in forest management planning.

Minimum harvesting age is an important parameter of forest management planning in deciding amount and variety of forest products can be harvested from forests. As known, the timing of interventions (i.e., minimum harvesting ages) to attain the best mix of forest products and services greatly affects the production of goods and services. For instance, Cooper (1983) showed the effects of rotation ages on carbon sequestered in trees and soil. Calish et al. (1978) examined the changes of optimal economic rotation ages when some forest resources such as water quantity and mass soil movement are integrated into forest management planning. While deer management shortens the rotation age of Douglas fir, for example, mass soil movement lengthens it. Recently, Diaz-Balteiro & Romero (2003), Backéus et al. (2005) and Baskent & Mumcu-Kucuker (2010) incorporated carbon sequestration and water production into forest management planning to understand the trade-offs among forest values. Financial worth of carbon would influence the optimal rotation ages too. Some studies explained that when carbon prices increased, the minimum harvesting age would increase, but when timber prices increased, the minimum harvesting age would decrease (Van Kooten et al., 1995). In addition, Olschewskia & Benitez (2010) demonstrated that cost of carbon have important effect on minimum harvesting ages that may cause a doubling of minimum harvesting age in contrary to optimum 15 years when considering just wood harvest. However, very few studies are involved in evaluating both the interactions between timber harvest and carbon storage and the effects of different minimum harvesting ages on dynamics of forest ecosystems under management (Swanson, 2009; Mumcu, 2007). Therefore, it has become necessary to explore the influence of forest ecosystems onto the carbon storage and water production by evaluating the effects of various harvesting ages on forest dynamics.

This study mainly presents a multiple-use forest management planning methodology accommodating carbon storage, water and timber production combined to form forest management strategies. In this context, carbon storage, water and timber production values were quantified based on forest biomass and twenty-four management strategies were established to maximize NPV of timber, water, carbon or all of them. The effects of three minimum harvesting ages on the level of management objectives were also examined. Management strategies include one forest value in objective function subject to desired level for other forest values as constraints and forest policies such as desired level of products, even flow timber products, and no restriction option. Forest management models developed under various planning strategies were solved by using modified linear programming (LP) approach over a planning horizon. Forest performance indicators such as NPV and amount of all forest values were used to understand forest dynamics created by the various management strategies.

In this study, timber, water and carbon values in association with their monetary values were incorporated successfully into a multi purposed forest management plan and the effects of different minimum harvesting ages on these values were explored. Twenty-four management alternatives with various objectives, minimum ages and constraints were developed and solved with LP for a real case study area.

In current study, 180-year minimum harvesting age was used in conservation areas and three different minimum ages such as 80, 100 and 120 years were tested in timber production areas. Shorter harvesting ages facilitated higher timber and its NPV over the planning horizon. A similar study demonstrated that shorter harvesting ages increased the amount of harvested timber (Liski et al., 2001). The same study showed that the mean net incomes of timber were the highest for Scots pine and Norway spruce was managed with 90-year and 60-year rotation age respectively. The study obviously showed that when the financial values of timber are integrated into the model, the model overlooks some expenses like reforestation. As expected, additional constraints into management planning cause loss of volume and NPV of timber products (Raymer et al., 2005; Baskent et al., 2008b).

In addition, when carbon values from all strategies with the same objective and constraints were compared, 80-year harvesting age caused higher carbon storage to take place. The mature initial age class structure in the planning unit seemed to be the most important factor causing for this response. As known, older stands sequester less carbon because of slower growth rate while regenerated stands developing in a regulated forest sequester more carbon (Jarvis et al., 2005). Similarly, Backéus et al. (2005) demonstrated that total carbon storage increases at a slower rate when a forest matures over time, implying less amount of carbon flux in a forest. Perez-Garcia et al. (2005) figured out that when carbon emissions are taken into account, shorter rotations incline to increase total carbon storage. Liski et al. (2001), however, showed that while the cumulative amount of carbon sequestered in Scots pine forest increased with increasing harvesting age it reduced in Norway spruce forest. Similarly, some researchers indicated that some silviculturel interventions such as reforestation (Krcmar et al., 2001; Baskent et al., 2008b) or afforestation (Kaul et al., 2009) particularly in early periods, cause high biomass and provide positive effects on the sequestrated carbon despite the certain amount of expenses of reforestation.

However, the results indicated that total water production and NPV in different minimum harvesting ages did not follow a systematic trend observed in timber and carbon values. As the water production is related to the basal area, strategies *W2 regenerated further areas in the early periods and accordingly most of the forest openings were reforested in the first period. In following periods, the renewed or reforested sub-compartments created further basal area, resulting in less quantity of water production. Similar correlations were detected by other researches (Brown et al., 2005; Benyon et al., 2007).

This study showed that the amount of forest products and services derived from forest ecosystem depend on minimum harvesting ages as well as forest management activities. Thus, the minimum harvesting age can then be considered a good tool and effective method to manage the forest products and services in developing forest management policies (Liski et al., 2001). Additionally, the performance of a management strategy depends highly on the components of a strategy and the initial forest structure aside from the growth rate.

The model developed in this study may have some shortcomings for further improvements. The stand simulation model is related to a simple allometric relationship between the current and the optimal development pattern of the stands. Realistically, however, a dynamic growth and yield model should be developed based on permanent sample plots and site information. There are many forest management objectives such as controlling soil loss that may have to be integrated into the model as well. The spatial arrangement of the harvest schedule such as block size, adjacency and opening sizes needs to be controlled either by using mixed integer programming or meta-heuristic techniques. Decomposition rates of timber should be calculated according to time and species too. The amount of soil carbon should also be taken into account in calculating total sequestered carbon. Furthermore, stochastic incidents such as forest fires, wind, insect fungi and climatic change affecting forest products and monetary values should also be involved in multiuse forest management planning.