Scenario-based estimation of catchment carbon storage: linking multi-objective land allocation with InVEST model in a mixed agriculture-forest landscape

Abstract

This study performed a scenario-based land allocation in a mixed agriculture-forest landscape in northern Iran to investigate how different land use policies contribute to changes in carbon storage. In pursuit of this goal, a temporal profile of the trade-off between the region’s land use land cover (LULC) classes was produced using Landsat image of the year 2016. The weighted linear combination procedure was also used to map the suitability of land for agriculture, forest, urban, and rangeland based on ecological and socio-economic criteria. The suitability maps were analyzed through the Multi-Objective Land Allocation procedure under five scenarios with differing areas devoted to each LULC to generate different patterns of LULC distribution in the region. In addition, the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model was used to estimate the potential of LULC classes in carbon storage. The amount of carbon storage differed significantly between the scenarios, ranging from 1.29 tons/ha/year when the majority of the land was devoted to agriculture (76% of the area) to 5.40 tons/ha/year when the landscape was dominated by forest (77% of the area). The extreme conditions presented in this research may not be as likely to occur, but opens a dialog between different stakeholders and informs of a probable trend of ecosystem service loss due to agricultural land expansion.

Keywords

multi-objective land allocation / carbon storage / InVEST model / Iran

Introduction

Carbon storage, in the context of ecosystem services, refers to the ability of terrestrial ecosystems in capturing and storing carbon in the form of soil organic matters as well as dead and live vegetation biomass (^{Boone et al., 2018}). Foremost among these forms is the aboveground live biomass that constitutes one of the Earth’s largest carbon pools, especially in forest areas (Mekonnen and Tolera, 2019). Carbon storage across large geographical extents, such as catchment area, is indicative of a regulating ecosystem service termed as carbon stock and carbon sequestration (^{Zhang et al., 2007}). The importance of this service is reflected in its role in the decarbonization of the biosphere and climate change adaptation by serving as the largest sink of atmospheric carbon (^{Lal et al., 2013}). Unfortunately, human beings tend to be attracted to areas supporting the growth of carbon-storing live vegetation biomass to meet their ever-increasing needs (^{de Souza Medeiros et al., 2020}). Enormous loss of forested areas and rangelands due to agricultural and urban land expansion has been the most adverse consequence of such a human desire. According to research conducted by ^{Hansen et al. (2013)}, a total of 2.3 million km² of forests were lost due to disturbance over the years 2000 to 2012. Such a rate of forest resource depletion is alarmingly high and undermines the role that they play in provision of ecosystem services and combating climate change (^{Hansen et al., 2013}).

Projections present a gloomy picture for the future of the Earth’s forests. Thanks to the progressive development of remote sensing (RS) and Geographic Information System (GIS) techniques, however, decision-makers and planners can now be informed regarding the consequences of their decisions upon land resources (^{Berg et al., 2016}). In other words, decision-makers have now the benefit of intuitively observing the effect of every single decision on the exchange and trade-off between different land use/land cover (LULC) types, especially loss of valuable land components in favor of urban and agricultural land expansion. Carbon cycling depends on LULC type with a various amount of CO₂ released to the atmosphere as a result of forest alteration to croplands and urban land use (^{Gao et al., 2013}). This ability has also made it possible to determine the impact of each envisaged decision upon land through scenario development perspectives (^{Thompson et al., 2016}; ^{Armenteras et al., 2019}; ^{Kucsicsa et al., 2019}).

Many important strides have also been made around the globe to develop methods that quantify ecosystems services (^{Turner et al., 2016}). The InVEST model (an acronym for Integrated Valuation of Environmental Services and Tradeoffs) is perhaps one of the most well-known and frequently applied techniques for ecosystem service assessment, especially carbon storage. InVEST software gives a spatially explicit view of the ecosystem service and incorporates user-defined preferences using scenario development and land restoration considerations (^{Sahle et al., 2019}). To address this limitation, researchers have shown a vested interest in the integration of ecosystem service assessment methods with scenario-based LULC change modeling and multi ceiteria and objective land allocation systems as a coupled framework. This enables them to understand how LULC decisions, leading to differing LULC areas and patterns and contributing to changes in ecosystem services (^{Kubiszewski et al., 2017}; ^{Zarandian et al., 2017}; ^{Asadolahi et al., 2018}).

Using historical RS-derived LULC maps, a pool of studies have dealt with the changes brought about by human intervention on ecosystem services. For example, adopting the InVEST model, ^{Cabral et al. (2016)} found a noticeable drop in various landscape ecosystem services due to urban expansion, especially carbon storage, in the urban community of Bordeaux, France over the period 1990–2006. ^{Pareta and Pareta (2011)} calibrated InVEST using a diversity of RS-based data to estimate carbon stocks in the Sagar district forest of India at a stand scale. In a step forward, the integration of Multi-Objective Land Allocation (MOLA) and LULC change models with InVEST has also gained recognition among researchers to foresee the results of various management practices for the future. For instance, ^{Liang et al. (2017)} developed a coupled framework by integrating CLUE-S (the Conversion of Land Use and its Effects at Small regional extent) with system dynamic and InVEST models to simulate and predict the pixel and regional-scale changes in carbon storage in the Zhangye oasis, northwest China. The simulated spatio-temporal expansion of urban areas under three scenarios (historical demand, moderate and strict protection) were used as inputs to InVEST, with results indicating notable changes in carbon storage under different probable decisions. With the same spirit, ^{He et al. (2016)} linked the Land Use Scenario Dynamics-urban (LUSD-urban) and InVEST models to leverage their combined advantage for estimating carbon storage under various change scenarios in Beijing and provided suggestions for more environmentally sound future urban expansion. The same methodologies with more promising results can be also found in other insightful studies like ^{Jiang et al. (2017)}, ^{Lyu et al. (2019)}, and ^{Sallustio et al. (2015)}.

In the wake of these studies, an attempt was made herein to investigate the impact of a number of land decisions on LULC pattern and carbon storage in Lahijan Catchment, northern Iran where the region is renowned for its historical Caspian Hyrcanian mixed forests and the rapidity at which urban and agricultural activities are changing the face of the area. In pursuit of this goal, land use maps of the region were delineated from Landsat images and applied for multi-objective land allocation in the study area under three different scenarios. The InVEST model was used afterward to determine how different LULC allocations contribute to carbon storage, thus suggesting useful storylines for future development in the region.

Materials and methods

Study area

Lahijan Catchment in Guilan Province, Northern Iran was selected as the study area (Fig. 1). With 13 sub-catchments, the region has an aerial extent of 542.5 km² spanning between 49°45′ to 50°13′€E longitude and 37°05′ to 37°27′€N latitude. The topography of the catchment is mountainous at the upstream in the south and plain at the downstream in the north. The upstream highlands are primarily pristine and are covered with dense Caspian Hyrcanian mixed forests, accounting for 55% of Lahijan while low-laying northern areas facing the Caspian Sea are dominated by agricultural activities and disperse farming communities. As the rainiest part of the country, the mean annual precipitation of the catchment is above 1350 mm. The region is also characterized by a Mediterranean climate according to the De Martonne’s climatic classification scheme with hot and humid summers and cold winters (^{Khormali et al., 2012}). According to Iranian census data collected in 2016, the permanent residents of the catchment are estimated to be around 700 thousand people, of which 76% live in urbanized settings. The majority of these people engage in farming activities, especially rice cultivation. LULC change in the region has escalated during the past four decades due mainly to increasing population growth which indicates the adoption of more environmentally sound LULC decisions to strike a balance between the pristine forested and intense agricultural and urbanized parts of the catchment. Figure 1 shows the layout of Lahijan Catchment in northern Iran.

Methods

LULC mapping

Three Landsat images covering the catchment were acquired from the Earth Explorer website for the years 1984 from the Thematic Mapper (TM) sensor, 2000 from the Enhanced Thematic Mapper Plus (ETM+ ) sensor, and 2016 from the Operational Land Imager (OLI) sensor. The images were classified using the supervised algorithm of Maximum Likelihood (^{Otukei and Blaschke, 2010}) into five classes of urban, forest, agriculture, rangeland, and water. The accuracy of the classification procedure was tested by building a confusion matrix for each LULC map using reference points collected through field surveys for 2016 and through the interpretation of false color composite images for 2000 and 1984.

LULC suitability evaluation

In this section, a land suitability map was produced for each LULC using the Weighted Linear Combination (WLC) procedure including urban, agriculture, rangeland, and forest (except for water). This procedure combines a number of spatial factors to determine the suitability of land for a target LULC (^{Malczewski, 2000}). In doing so, factors should be weighted relatively based on their level of contribution and standardized independently prior to combination. The analytical hierarchy process (AHP) (^{Vahidnia et al., 2008}) and fuzzy logic (^{Chang et al., 2008}) are among the most preferred techniques in the literature to carry out these analyses. AHP is a simple but potent method capable of eliciting criteria weights with respect to other factors incorporated in the weighting procedure. Technically, AHP performs a pairwise comparison between every two criteria using a point scale to convey individual judgments (^{De Felice et al., 2016}). Moreover, fuzzy logic has proved effective in the standardization of input layers in the WLC procedure (^{Kuznichenko et al., 2019}). In this process, continuous raster layers are fuzzified using fuzzy membership functions while categorical maps are often standardized using rating scales. As the WLC procedure was conducted for each LULC type independently to perform the multi-objective allocation task, the experts involved in the AHP weighting analysis were recruited based on their knowledge about the type of LULC asked as well as the socio-economic and biophysical conditions of the region. The factors used in the WLC procedure for each LULC suitability evaluation together with the weights derived from AHP and fuzzification method performed on them are given in Table 1.

Multi-objective land allocation

The multi-objective land allocation (MOLA) module in Idrisi Selva was used to assign a proper use of land to each part of the catchment (i.e., pixels) subjecting to the constraints applied under five scenarios. This module is developed to solve multi-objective land allocation issues facing conflicting objectives by seeking an optimal compromised solution based on an iterative process in which the suitability of land is maximized for all LULCs regarding their ranking and weighting (^{Irina et al., 2019}). In other words, MOLA tries to allocate each unit of land to a LULC type to which the total weighted suitability is as closest to the optimum (i.e., the highest-ranking weight). Having produced a land suitability map for each LULC, the AHP procedure was used to assign a relative weight to the LULC classes of interest. The rank module in Idrisi was also employed to re-scale the suitability maps in ascending order by assigning 1 to the highest suitability value.

Five scenarios were developed by changing the area assigned to each LULC type for allocation using MOLA. In scenario S0, the area assigned to each LULC was equal to that of the 2016 Landsat-derived LULC map and considered as the baseline scenario (S0). In scenario S1, the area given to each LULC type was considered as constant except for rangeland whose area was assigned according to its ecological suitability. Scenario S2 was developed by duplicating the area given to agricultural areas. Scenario S3 was devised by increasing the quota of forest area according to its respective suitability map. Scenario S4 was developed based on probable LULC changes that may occur in the area in the near future. Future LULC changes were determined using the Markov chain analysis in Idrisi. The 1984 and 2016 LULC maps were introduced to the Markov chain module to estimate the possible transition area between LULC classes up to 2030 and the resulting areas were used to devise scenario S4. The MOLA ultimately combined the rank maps according to their weights and the scenario areas to generate a LULC map under each scenario.

Determining carbon storage using InVEST model

InVEST uses the following equation (Eq. (1); ^{Sharp et al., (2014)}) to estimate carbon storage where C_xt is the amount of carbon stored in a unit of area (pixel) at time t which is equal to the sum of carbon stored in the aboveground (C_aj) and below-ground vegetation(C_bj), soil (C_sj), and dead organic matter (C_oj) per ha (Mg of C ha-1), J is the type of LULC, A_xjt is the area of LULC j in pixel x at time t, and C_pxt is the amount of harvested wood products (HWPs) in pixel x at time t. The unknown parameters of this equation were derived from the Intergovernmental Panel on Climate Change (IPCC) data sets as suggested by ^{Asadolahi et al. (2017)}.

(1){\mathrm{C}}_{\mathit{xt}}{\mathrm{=C}}_{\mathit{pxt}}\mathrm{+}{\sum }_{j\mathrm{=1}}^{\mathrm{J}}{\mathrm{A}}_{\mathit{xjt}}{\mathrm{(C}}_{\mathrm{a}j}{\mathrm{+C}}_{\mathrm{b}j}{\mathrm{+C}}_{\mathrm{s}j}{\mathrm{+C}}_{\mathrm{o}j}).

Assessing the impact of different LULC allocation on carbon storage

Having produced different LULC allocation under five different scenarios and the amount of carbon that can be stored potentially per unit of area, the Zonal Statistic was finally employed to investigate how different combinations and patterns of LULC classes contribute to the amount of carbon stored across the understudy catchment.

Results

The LULC maps of the region, illustrated in Fig. 2, were generated with Kappa values ranging from 0.79 to 0.86 (Table 2). Forest, followed by agriculture was characterized as the predominant LULC classes of the catchment. The area of forest dropped from 69.1% in 1984 to below 53% in 2016 while the percentage of area occupied by urban, agriculture and rangeland increased by 3.75%, 0.33%, and 12.4% during the study period, respectively (Table 2). The majority of the LULC changes in the region were associated with forest loss due to agricultural expansion in northern parts of the area.

The resulting maps of land suitability evaluation for different LULC classes are presented in Fig. 3. The most suitable areas for forest were located between an altitudinal range of 0 and 1000 m, slope range of 0 and 25% on deep and over medium-high drainage soils in the south. Agricultural areas preferred areas in the north where soil is relatively deep and well-structured with a good humus horizon and loam clay texture. The suitability of land for urban had a similar distribution pattern to that of agriculture with the most suitable areas located in low-lying areas with a slope range of below 9 degrees sufficiently away from main streams as well as marl and schist parental rock. The region had relatively low suitability for rangeland according to the selected factors (see Table 1). The most suitable rangelands, however, were distributed in areas having an altitudinal range of 0–1000 m and a slope of 0–15% as well as areas free of forest, urban and agriculture.

The results of scenario-based multi-objective land allocation are shown in Fig. 4 and represented in Table 3. The area allocated to urban remained relatively constant between scenarios, ranging from 2892 ha (in S4) to 3032 ha (in S1). In scenarios S1 to S3, the area of rangeland was as low as 86 ha while an area of 554 ha was allocated to this LULC in scenario S4. In scenario S2, the proportion of area allocated to agriculture increased largely (76% of the total area) at the expense of decreasing forest land to nearly 17%. Contrarily, forest achieved the largest area quota of 77% in scenario S3 in which the area allocated to agriculture was as low as 16% of the land.

The results of InVEST carbon storage model under five different scenarios of LULC allocation are given in Fig. 5 and Table 4. The southern forest dominated sub- catchments yielded higher carbon storage values while northern sub-basins where the landscape is mostly covered with urban and agriculture had relatively lower values of carbon storage (tons/ha/yr). Scenarios S0, S1, and S4 performed somehow similarly in storing carbon (2.48 to 2.90 tons/ha/yr). In scenario S2, in which the largest proportion of land was allocated to agriculture, the amount of carbon storage was the lowest among the scenarios with only two sub-basins having carbon storage values of above 3 tons/ha/yr. Scenario S3, however, outperformed others in capturing carbon with a mean storage value of 5.4 tons/ha/yr.

Discussion and conclusions

Natural vegetation covers play a leading role in the Earth’s carbon equilibrium and abatement of greenhouse gas accumulation. While these roles, together with the basic principles underlying the contribution of LULC management to carbon storage and emission, have been well articulated and acknowledged in the literature, effective incorporation of carbon management into land use policies and management still remains a tantalizing task (^{Fahey et al., 2010}). Thanks to the progressing development of dynamic GIS and RS techniques, a reliable foundation has been laid to realistic assessment of organic stock and inform the LULC contribution to carbon sequestration under the influence of several factors such as climate change (^{Nelson et al., 2010}) and local synergies and tradeoffs (^{Grêt-Regamey et al., 2013}). Particularly, integration of LULC change and allocation models with methods quantifying the exchange of carbon between terrestrial ecosystems and the atmosphere has gained much popularity in recent years and opened promising perspective for better management and protection of land under the unprecedented and rapid human-centered development in many parts of the world, especially in developing countries (^{He et al., 2016}).

A case study in Lahijan Catchment, northern Iran showed that different perspectives and prioritization of land use activities would lead to a considerable variation in the amount of carbon (nearly 4.11 ton/ha/year) that could be sequestrated in a region with a mixed agriculture-forest background matrix. More evidences have been found in similar studies, demonstrating that any decision made over land would significantly affect the carbon sequestration functioning of the landscape. For instance, the findings of ^{Jiang et al. (2017)} indicated that the ability of a mixed landscape in capturing atmospheric carbon varies significantly by 1.52 Tg when adopting different urban growth policies over the course of nearly a decade. Contrary to the majority of studies implicating different urban expansion policies in lowering the ability of the landscape to capture and store carbon (see the literature review provided by ^{Liang et al. (2017)} and ^{Jiang et al. (2017)}), our findings showed that, at regional scales, significant variation in carbon sequestration rested more on different quota of land shared between forest and agriculture under different scenarios than differing expansion and distribution of urban areas that constituted a small proportion of the region.

Considering the high probability of forest conversion to agriculture and vice versa, the scenarios constructed in the present study aimed to portray two extremes of land allocation between agriculture and forest (scenarios S2 and S3) that, according to the history of the region (see Fig. 2), are more likely to be converted to each other in the future. Of note is that it is highly unlikely to expect the occurrence of these two extreme scenarios in the region (except for scenarios S0, S1, and S4), but their results are representative of the maximum and minimum capability of region to contribute to carbon storage and are also informative in deciding which combination of LULC classes, especially agriculture and forest, is an optimal compromise solution for the region. As shown in Fig. 1 and Fig. 2, the area occupied by these two particular LULC types is a function of topography, but their main bordering line in the middle of the region has been moving quickly in a southerly direction (especially from 1984 onwards) to create more open spaces for agriculture. Hence, more realistic and applicable scenarios should be more attentive to the changes that the border may face in the future and the ensuing consequences on carbon sequestration.

While the results of the present study provide useful implications for LULC policy decision making, more research is still required to take a broader range of ecosystem services into account. Moreover, a thorough investigation of LULC change, simulation, and allocation adopting more sophisticated techniques should be part of the coupled LULC-ecosystem services modeling framework to make robust decisions on land, especially in terms of socio-economic characteristics (^{Liu et al., 2017}). The InVEST model has also been a common part of this framework in the literature and also in the present study due to its simplicity of application and calculation. However, as Tallis et al. (2013) imply, an important issue that must be taken fully into consideration when interpreting InVEST results is that it focuses on carbon density rather than biogeochemical attributes that are a major part of the carbon cycle. Additionally, seasonal variation in the foliage of deciduous Caspian Hyrcanian mixed forests merits further attention in the region to reach a better pattern of LULC allocation in the region.

Although this study had no intention of feeding current land use planning, it opens a dialog between different stockholders of environmental protection and agricultural demands, and motivates a strong call for the integration of ecosystem services in land use planning and management in northern Iran where the vested desire of locals in agricultural and farming activities have caused significant loss of Caspian Hyrcanian mixed forests and the ecosystem services that they bring to local and global communities. The extreme conditions presented by the coupled InVEST-MOLA procedure in this research may not be as likely to occur, but inform of a continuous process aiming to open further areas for agriculture at the expense of valuable forest areas.

References

[1] Armenteras D, Murcia U, González T M, Barón O J, Arias J E (2019). Scenarios of land use and land cover change for NW Amazonia: impact on forest intactness. Glob Ecol Conserv, 17: e00567

[2] Asadolahi Z, Salmanmahiny A, Sakieh Y (2017). Hyrcanian forests conservation based on ecosystem services approach. Environ Earth Sci, 76(10): 365

[3] Asadolahi Z, Salmanmahiny A, Sakieh Y, Mirkarimi S H, Baral H, Azimi M (2018). Dynamic trade-off analysis of multiple ecosystem services under land use change scenarios: towards putting ecosystem services into planning in Iran. Ecol Complex, 36: 250–260

[4] Berg C, Rogers S, Mineau M (2016). Building scenarios for ecosystem services tools: developing a methodology for efficient engagement with expert stakeholders. Futures, 81: 68–80

[5] Boone R B, Conant R T, Sircely J, Thornton P K, Herrero M (2018). Climate change impacts on selected global rangeland ecosystem services. Glob Change Biol, 24(3): 1382–1393

[6] Cabral P, Feger C, Levrel H, Chambolle M, Basque D (2016). Assessing the impact of land-cover changes on ecosystem services: a first step toward integrative planning in Bordeaux, France. Ecosyst Serv, 22: 318–327

[7] Chang N B, Parvathinathan G, Breeden J B (2008). Combining GIS with fuzzy multicriteria decision-making for landfill siting in a fast-growing urban region. J Environ Manage, 87(1): 139–153

[8] De Felice F, Petrillo A, Saaty T (2016). Applications and Theory of Analytic Hierarchy Process: Decision Making for Strategic Decisions—Application of the Ahp Method in Environmental Engineering: Three Case Studies. Paris: BoD–Books on Demand

[9] de Souza Medeiros A, Malta Ferreira Maia S, dos Santos T C, de Araújo Gomes T C (2020). Soil carbon losses in conventional farming systems due to land-use change in the Brazilian semi-arid region. Agric Ecosyst Environ, 287: 106690

[10] Fahey T J, Woodbury P B, Battles J J, Goodale C L, Hamburg S P, Ollinger S V, Woodall C W (2010). Forest carbon storage: ecology, management, and policy. Front Ecol Environ, 8(5): 245–252

[11] Grêt-Regamey A, Celio E, Klein T M, Wissen Hayek U (2013). Understanding ecosystem services trade-offs with interactive procedural modeling for sustainable urban planning. Landsc Urban Plan, 109(1): 107–116

[12] Gao Z, Cao X, Gao W (2013). The spatio-temporal responses of the carbon cycle to climate and land use/land cover changes between 1981–2000 in China. Front Earth Sci, 7(1): 92–102

[13] Hansen M C, Potapov P V, Moore R, Hancher M, Turubanova S A, Tyukavina A, Thau D, Stehman S V, Goetz S J, Loveland T R,Kommareddy A (2013). High-resolution global maps of 21st-century forest cover change. Science, 342(6160): 850–853

[14] He C, Zhang D, Huang Q, Zhao Y (2016). Assessing the potential impacts of urban expansion on regional carbon storage by linking the LUSD-urban and InVEST models. Environ Model Softw, 75: 44–58

[15] Irina L T, Javier B P, Teresa C B M, Eurídice L A, Luz María del Carmen C I (2019). Integrating ecological and socioeconomic criteria in a GIS-based multicriteria-multiobjective analysis to develop sustainable harvesting strategies for Mexican oregano Lippia graveolens Kunth, a non-timber forest product. Land Use Policy, 81: 668–679

[16] Jiang W, Deng Y, Tang Z, Lei X, Chen Z (2017). Modelling the potential impacts of urban ecosystem changes on carbon storage under different scenarios by linking the CLUE-S and the InVEST models. Ecol Modell, 345: 30–40

[17] Khormali F, Ayoubi S, Kananro Foomani F, Fatemi A (2012). Tea yield and soil properties as affected by slope position and aspect in Lahijan area, Iran. Int J Plant Prod, 1(1): 99–111

[18] Kubiszewski I, Costanza R, Anderson S, Sutton P (2017). The future value of ecosystem services: global scenarios and national implications. Ecosyst Serv, 26: 289–301

[19] Kucsicsa G, Popovici E A, Bălteanu D, Dumitraşcu M, Grigorescu I, Mitrică B (2019). Assessing the potential future forest-cover change in Romania, predicted using a scenario-based modelling. Environ Model Assess, 25:471–491.

[20] Kuznichenko S, Buchynska I, Kovalenko L, Gunchenko Y (2019). Suitable site selection using two-dtage GIS-based fuzzy multi-criteria decision analysis. In: International Conference on Computer Science and Information Technology. Berlin: Springer, Cham: 214–230

[21] Lal R, Lorenz K, Hüttl R F, Schneider B U, Von Braun J (2013). Ecosystem Services and Carbon Sequestration in the Biosphere. Dordrecht: Springer

[22] Liang Y, Liu L, Huang J (2017). Integrating the SD-CLUE-S and InVEST models into assessment of oasis carbon storage in northwestern China. PLoS One, 12(2): e0172494

[23] Liu J, Zhang G, Zhuang Z, Cheng Q, Gao Y, Chen T, Huang Q, Xu L, Chen D (2017). A new perspective for urban development boundary delineation based on SLEUTH-InVEST model. Habitat Int, 70: 13–23

[24] Lyu R, Mi L, Zhang J, Xu M, Li J (2019). Modeling the effects of urban expansion on regional carbon storage by coupling SLEUTH-3r model and InVEST model. Ecol Res, 34(3): 380–393

[25] Malczewski J (2000). On the use of weighted linear combination method in GIS: common and best practice approaches. Transactions in GIS, 4(1): 5–22

[26] Mekonnen A, Tolera M (2019), Carbon stock estimation along altitudinal gradient in Sekele-Mariam dry evergreen montane forest, north-western Ethiopia agriculture. Forestry and Fisheries, 8: 48

[27] Nelson E, Sander H, Hawthorne P, Conte M, Ennaanay D, Wolny S, Manson S, Polasky S (2010). Projecting global land-use change and its effect on ecosystem service provision and biodiversity with simple models. PLoS One, 5(12): e14327

[28] Otukei J R, Blaschke T (2010). Land cover change assessment using decision trees, support vector machines and maximum likelihood classification algorithms. Int J Appl Earth Obs Geoinf, 12: S27–S31

[29] Pareta K, Pareta U (2011). Forest carbon management using satellite remote sensing techniques: a case study of Sagar district (MP). International Scientific Research Journal, 3(4): 335–348

[30] Sahle M, Saito O, Fürst C, Demissew S, Yeshitela K (2019). Future land use management effects on ecosystem services under different scenarios in the Wabe River catchment of Gurage Mountain chain landscape, Ethiopia. Sustain Sci, 14(1): 175–190

[31] Sallustio L, Quatrini V, Geneletti D, Corona P, Marchetti M (2015). Assessing land take by urban development and its impact on carbon storage: findings from two case studies in Italy. Environmental Impact Assessment Review, 54: 80–90

[32] Senior RA, Hill JK, Edwards DP (2019). Global loss of climate connectivity in tropical forests. Nat Clim Chang, 9: 623–626

[33] Sharp R, Chaplin-Kramer R, Wood S, Guerry A, Tallis H, Ricketts T H (2014). InVEST user’s guide: integrated valuation of environmental services and tradeoffs. In: Stanford Woods Institute for the Environment. University of Minnesota’s Institute on the Environment, The Nature Conservancy & WW Foundation Stanford

[34] Tallis H T, Ricketts T, Guerry A D, Wood S A, Sharp R, Nelson E, Ennaanay D, Wolny S, Olwero N, Vigerstol K, Pennington D (2013). InVEST 2.5. 3 User’s Guide. The natural capital project. Stanford, (Available at Stanford website, last accessed February, 2020)

[35] Thompson J R, Lambert K F, Foster D R, Broadbent E N, Blumstein M, Almeyda Zambrano A M, Fan Y (2016). The consequences of four land-use scenarios for forest ecosystems and the services they provide. Ecosphere, 7(10): e01469

[36] Turner K G, Anderson S, Gonzales-Chang M, Costanza R, Courville S, Dalgaard T, Dominati E, Kubiszewski I, Ogilvy S, Porfirio L, Ratna N, Sandhu H, Sutton P C, Svenning J C, Turner G M, Varennes Y D, Voinov A, Wratten S (2016). A review of methods, data, and models to assess changes in the value of ecosystem services from land degradation and restoration. Ecol Modell, 319: 190–207

[37] Vahidnia M H, Alesheikh A, Alimohammadi A, Bassiri A (2008). Fuzzy analytical hierarchy process in GIS application. Int Arch Photogramm Remote Sens Spat Inf Sci, 37(B2): 593–596

[38] Zarandian A, Baral H, Stork N E, Ling M A, Yavari A R, Jafari H R, Amirnejad H (2017). Modeling of ecosystem services informs spatial planning in lands adjacent to the Sarvelat and Javaherdasht protected area in northern Iran. Land Use Policy, 61: 487–500

[39] Zhang J, Ge Y, Chang J, Jiang B, Jiang H, Peng C, Zhu J, Yuan W, Qi L, Yu S (2007). Carbon storage by ecological service forests in Zhejiang Province, subtropical China. For Ecol Manage, 245(1–3): 64–75

RIGHTS & PERMISSIONS

Higher Education Press