No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Transition potential modeling for land-cover change
... Spatially-explicit, pattern-based, inductive models of land change rely on mathematical representations of the relationship between independent variables and past land change to extrapolate future land change [2,3]. Machine learning techniques that quantify these mathematical relationships have become popular compared to older parametric techniques, such as logistic regression [4]. ...
... Population Empirical Probability for category k = size of change on category k size of category k Eastman et al. (2005) state that a potential benefit of transforming categorical variables to PEL is that the calculated relationships are independent of the size of each category, and therefore PEL are transferrable across time and space, unlike PEP [4]. The benefit of transforming categorical variables as PEL was a speculation by the authors, who concluded PEL would probably outperform PEP. ...
... Our results illustrate that SEP encoding and binary encoding produce similar output values, which rank the categories consistently in sequence of change intensity. Eastman et al. (2005) proposed PEL to encode a categorical variable while avoiding the creation of a collection of binary variables [4]. SEP meets this same desirable property, which conserves computer resources. ...
The profession debates how to encode a categorical variable for input to machine learning algorithms, such as neural networks. A conventional approach is to convert a categorical variable into a collection of binary variables, which causes a burdensome number of correlated variables. TerrSet’s Land Change Modeler proposes encoding a categorical variable onto the continuous closed interval from 0 to 1 based on each category’s Population Evidence Likelihood (PEL) for input to the Multi-Layer Perceptron, which is a type of neural network. We designed examples to test the wisdom of these encodings. The results show that encoding a categorical variable based on each category’s Sample Empirical Probability (SEP) produces results similar to binary encoding and superior to PEL encoding. The Multi-Layer Perceptron’s sigmoidal smoothing function can cause PEL encoding to produce nonsensical results, while SEP encoding produces straightforward results. We reveal the encoding methods by illustrating how a dependent variable gains across an independent variable that has four categories. The results show that PEL can differ substantially from SEP in ways that have important implications for practical extrapolations. If users must encode a categorical variable for input to a neural network, then we recommend SEP encoding, because SEP efficiently produces outputs that make sense.
... The remote sensing-based approach implemented here is based on IDRISI software's LCM module [20,21]. Modeling changes in LULC via the LCM is performed in five main stages: (i) historical change analysis; (ii) modeling of potential transition (At this stage, the probability of transfer from one type of land cover to another was determined by stimulus variables) [21]; (iii) LULC change prediction (2010); (iv) assessing model accuracy; and (v) LULC change projection for the future (2040). ...
... The remote sensing-based approach implemented here is based on IDRISI software's LCM module [20,21]. Modeling changes in LULC via the LCM is performed in five main stages: (i) historical change analysis; (ii) modeling of potential transition (At this stage, the probability of transfer from one type of land cover to another was determined by stimulus variables) [21]; (iii) LULC change prediction (2010); (iv) assessing model accuracy; and (v) LULC change projection for the future (2040). For the case study, LULC maps for 1984 and 2001, and five spatially-distributed variables [21]-normalized difference vegetation index (NDVI), distance to rivers, distance to the edge of forests, distance to pasture lands and empirical likelihood to change (To produce the empirical likelihood to change (qualitative variable), a transition map was generated from all classes to agriculture)-were used to model the transition potential (i.e., the potential for a LULC to transition to another LULC). ...
... Modeling changes in LULC via the LCM is performed in five main stages: (i) historical change analysis; (ii) modeling of potential transition (At this stage, the probability of transfer from one type of land cover to another was determined by stimulus variables) [21]; (iii) LULC change prediction (2010); (iv) assessing model accuracy; and (v) LULC change projection for the future (2040). For the case study, LULC maps for 1984 and 2001, and five spatially-distributed variables [21]-normalized difference vegetation index (NDVI), distance to rivers, distance to the edge of forests, distance to pasture lands and empirical likelihood to change (To produce the empirical likelihood to change (qualitative variable), a transition map was generated from all classes to agriculture)-were used to model the transition potential (i.e., the potential for a LULC to transition to another LULC). Transition potential included forest to agriculture, forest to pasture and pasture to agriculture. ...
Non-point source pollution is a major factor in excessive nutrient pollution that can result in the eutrophication. Land use/land cover (LULC) change, as a result of urbanization and agricultural intensification (e.g., increase in the consumption of fertilizers), can intensify this pollution. An informed LULC planning needs to consider the negative impacts of such anthropogenic activities to minimize the impact on water resources. The objective of this study was to inform future land use planning by considering nutrient reduction goals. We modeled the LULC dynamics and determined the capacity for future agricultural development by considering its impacts on nitrate runoff at a watershed scale in the Tajan River Watershed in northeastern Iran. We used the Soil and Water Assessment Tool (SWAT) to simulate the in-stream nitrate concentration on a monthly timescale in this watershed. Historical LULCs (years 1984, 2001 and 2010) were derived via remote sensing and were applied within the Land Change Modeler to project future LULC in 2040 under a business-as-usual scenario. To reduce nitrate pollution in the watershed and ecological protection, a conservation scenario was developed using a multi-criteria evaluation method. The results indicated that the implementation of the conservation scenario can substantially reduce the nitrate runoff (up to 72%) compared to the business-as-usual scenario. These results can potentially inform regional policy makers in strategic LULC planning and minimizing the impact of nitrate pollution on watersheds. The proposed approach can be used in other watersheds for informed land use planning by considering nutrient reduction goals.
... The most common methods used in land cover prediction modeling are weights of evidence [18], logistic regression [14,22], and multi-layer perceptron (MLP) neural networks [23]. Recent research has shown that MLP neural networks perform better than other methods in land change modeling since they can more adequately process non-linear relationships [11,24,25]. The algorithm for the MLP neural network is non-parametric and does not consider multicollinearity. ...
... In Step 3, transition potential maps were created to represent the likelihood for a patch of mature forest to be converted to an anthropogenic patch [24]. These were generated using data from the forest cover change analysis and spatial variables that had been selected in step two as the drivers of deforestation in the MGL. ...
... Transition potentials can be developed with several methods including logistic regression [14,22], multi-layer perceptron (MLP) neural networks [23], and weights of evidence [18]. This model was implemented with a MLP neural network, as recent studies have found that it outperforms other methods [11,24,25]. This study achieved an accuracy training rate of 82%, which is above the minimum required level. ...
Tropical forests and the biodiversity they contain are declining at an alarming rate throughout the world. Although southern Belize is generally recognized as a highly forested landscape, it is becoming increasingly threatened by unsustainable agricultural practices. Deforestation data allow forest managers to efficiently allocate resources and inform decisions for proper conservation and management. This study utilized satellite imagery to analyze recent forest cover and deforestation in southern Belize to model vulnerability and identify the areas that are the most susceptible to future forest loss. A forest cover change analysis was conducted in Google Earth Engine using a supervised classification of Landsat 8 imagery with ground-truthed land cover points as training data. A multi-layer perceptron neural network model was performed to predict the potential spatial patterns and magnitude of forest loss based on the regional drivers of deforestation. The assessment indicates that the agricultural frontier will continue to expand into recently untouched forests, predicting a decrease from 75.0% mature forest cover in 2016 to 71.9% in 2026. This study represents the most up-to-date assessment of forest cover and the first vulnerability and prediction assessment in southern Belize with immediate applications in conservation planning, monitoring, and management.
... Given a set of transition potentials, ELCM can then focus on determining the quantity of expected change (such as through Markov chain analysis) and the specific allocation, based on the transition potentials, to produce a forecast land cover map (Eastman et al. 2005). However, the most intensive part of the analysis is typically the development of the transition potentials. ...
... In case-control sampling, the prior probability is effectively 0.5. Most land change models use a greedy selection process where it is assumed that the pixels that should be changed in the prediction are those with the highest transition potential (Eastman et al. 2005). In cases such as this, no prior correction is needed since the transition potentials are monotonically related to those with prior correction. ...
... MLP neural networks have received considerable interest because of their ability to model complex non-linear relationships and their non-parametric character (Tso and Mather 2001;Lin et al. 2011). Furthermore, in testing a wide variety of approaches to the modeling of transition potentials, Eastman et al. (2005) and Lin et al. (2011) found MLP to be superior to other techniques tested in the skill of the prediction. The research questions guiding the evaluation were thus: ...
A critical foundation for empirical land change modeling is the mapping of transition potentials—quantitative evaluations of the readiness or suitability of land to go through a transition. This paper presents a procedure based on empirically determined normalized likelihoods of transition. It shows that these normalized likelihoods equate to posterior probabilities if case–control sampling is carried out among historical instances of change and persistence. The posterior probabilities can then be aggregated at the pixel level across multiple covariates using linear opinion pooling where the pixel-specific weight for each covariate is determined locally by its ability to distinguish between the alternatives of change or persistence. Thus, covariates with spatially varying diagnostic ability can be productively incorporated. The resulting algorithm is shown to have a skill comparable to that of a multi-layer perceptron approach with the advantage of high efficiency and amenability to distributed processing in a cloud environment.
... The land change modeler is a software for creating sustainable ecological development, which was designed to understand and identify land cover changes and environmental and protection requirements caused by these changes. This software exists as a vertical application in the IDRISI software system [28]. This model is also available as Extension for ArcGIS software. ...
... In the present research, Skill Measure [28] and Accuracy Rate were used to analyze the sensitivity of the multilayer perceptron artificial neural network model. Skill Measure is a statistic to evaluate the ability of the model based on the validation data and measures the skill of the model to predict future changes based on the training data. ...
The change in forest cover plays a vital role in ecosystem services, atmospheric carbon balance, and, thus, climate change. In this study, land use maps for the periods 1984 and 2012, derived from Landsat TM satellite imagery, were used. The goal of this study is comparison of three procedures of artificial neural network, logistic regression, and similarity weighted instance-based learning (SIM Weight) to predict spatial trend of forest cover change. The SimWeight considers the nearest instances in the variable space, which are computed based on past changes and the relative importance of the driving variables. The LogReg approach, on the other hand, is a type of generalized linear model that assumes that the current land use pattern reflects the processes of land use in the past. Artificial Neural Network is a nonparametric algorithm that is capable of fitting complex nonlinear functions to find the relations between past changes and their driving variables. Such approaches are expected to produce better fitting between the change potential and their complex relationships with their driving variables. Artificial neural networks in comparison with logistic regression and SimWeight have higher accuracy and less error in modeling and predicting of forest changes.
... Furthermore, the use of a Multilayer Perceptron neural network (MLP) to model the various transitions has been found to provide valuable information about the contribution of explanatory variables (Eastman 2016). The combination of CA Markov processes with MLP has been reported to offer the strongest capabilities among other approaches (Eastman, Solorzano, and Fossen 2005;Eastman 2016). Considering the many advantages of the CA Markov models, an attempt was made to create, test, apply and validate a CA Markov model at the country scale incorporating as explanatory variables environmental and socioeconomic parameters and simulating a set of multiple transitions, for an area with a diverse and complex land cover pattern, like Greece. ...
... Those are used to model the transition process in the past. Various models can be applied to simulate the transitions of land, such as machine learning techniques which can handle highly nonlinear problems, like the Multi-Layer Perceptron (MLP) or the Similarity Weighted Instance-based Learning (Sangermano and Eastman 2010) or even logistic regression, but only MLP can group and model multiple transitions with a single sub-model (Eastman, Solorzano, and Van Fossen 2005). Next, Driver Variables (also known as explanatory or predictor variables) are identified and their predictive ability in the model is evaluated. ...
Land change modeling (LCM) is a complex GIS procedure aiming at predicting land cover changes in the future, contributing thus to the design of interventions that help maintain ecosystem services and mitigate climate change impacts. In the present work, the land change model for Greece, a typical Mediterranean country, has been developed, based on historical information from remotely sensed land cover data. Land cover types based on the International Geosphere-Biosphere Program (IGBP) classification were obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) land cover product, i.e. MCD12Q1, provided annually from 2001 to 2018 at a spatial resolution of 500 m. Initially, the dominant land cover changes and their driving variables for the decade of 2001 to 2011 were determined and the transition potential of land was mapped using a multi-layer perceptron (MLP) neural network. Four dominant land-cover transformations were found in Greece from 2001 to 2011, i.e. land transformation from Savannas to Woody Savannas, from Savannas to Grasslands, from Grasslands to Savannas, and from Croplands to Grasslands. Driving variables were found to be the Evidence Likelihood of Land Cover, i.e. the relative frequency with which different land cover categories occurred within the areas that transitioned, the Altitude as realized in the Digital Elevation Model of Greece from ASTER GDEM, the Distance from previously changed land and two climate variables i.e. Mean Annual Precipitation and Mean Annual Minimum Temperature. After the model was calibrated, its predictive ability was tested for land cover prediction for 2018 and was found to be 96.7%. Future land cover projections up to 2030 were developed incorporating CMIP6 climate data under two Shared Socioeconomic Pathways (SSPs), i.e SSP126 corresponding to a sustainable future and SSP585, which describes the future world based on fossil-fueled development. The results indicate that major historical land transformations in Greece, do not correspond to land degradation or desertification, as it has been reported in previous works. On the contrary, the land cover transitions indicate that the Woody Savannas gain areas constantly, whereas Grasslands and Croplands lose areas, and forested areas of all types demonstrate moderate gains. Concerning future land cover, the present work indicates that the direction of historical changes will also prevail in the next decade, with the most severe scenario, i.e. SSP585 slowing down the rate of changes and the most sustainable one, i.e. SSP126, accelerating the rate of expansion of woody vegetation land cover type.
... According to [24], modeling the timing and the system changes tendency is crucial because of non-stationarity and systemic changes in land-use change projections. These changing drivers and processes influence the calibration [10,23,[25][26][27] and outcomes of empirical models [24,25,28,29]. In addition to that, it is challenging for available algorithms to capture the dynamic and nonlinear humanenvironment processes that drive complex land changes [5,23,26]. ...
... According to [24], modeling the timing and the system changes tendency is crucial because of non-stationarity and systemic changes in land-use change projections. These changing drivers and processes influence the calibration [10,23,[25][26][27] and outcomes of empirical models [24,25,28,29]. In addition to that, it is challenging for available algorithms to capture the dynamic and nonlinear humanenvironment processes that drive complex land changes [5,23,26]. ...
The temporal non-stationarity of land use and cover change (LUCC) processes is one of the main sources of uncertainty that may influence the calibration and the validation of spatial path-dependent LUCC models. In relation to that, this research aims to investigate the influence of the temporal non-stationarity of land change on urban growth modeling accuracy based on an empirical approach that uses past LUCC. Accordingly, the urban development in Rennes Metropolitan (France) was simulated using fifteen past calibration intervals which are set from six training dates. The study used Idrisi’s Cellular Automata-Markov model (CA-Markov) which is an inductive pattern-based LUCC software package. The land demand for the simulation year was estimated using the Markov Chain method. Model validation was carried out by assessing the quantity of change, allocation, and spatial patterns accuracy. The quantity disagreement was analyzed by taking into consideration the temporal non-stationarity of change rate over the calibration and the prediction intervals, the model ability to reproduce the past amount of change in the future, and the time duration of the prediction interval. The results show that the calibration interval significantly influenced the amount and the spatial allocation of the estimated change. In addition to that, the spatial allocation of change using CA-Markov depended highly on the basis land cover image rather than the observed transition during the calibration period. Therefore, this study provides useful insights on the role of the training dates in the simulation of non-stationary LUCC.
... Automatic machine learning was used for projecting future land cover maps. For this task, many different methods are described in the literature (e.g., logistic regression, similarity-weighted instance-based machine learning), but neural networks are the most commonly used and are also known to outperform the other methods [44,45]. For this simulation, we used the land change modeler module of TerrSet. ...
... Once trained, the neural networks were used to generate a set of land cover transition/persistence potentials [44]. Finally, using the past land cover changes and the modelled transition/persistence potentials, the future land cover map was predicted with a Markov Chain. ...
Erosion is one of the major threats listed in the Soil Thematic Strategy of the European Commission and the Alps are one of the most vulnerable ecosystems, with one of the highest erosion rates of the whole European Union. This is the first study investigating the future scenarios of soil erosion in Val Camonica and Lake Iseo, which is one of the largest valleys of the central Italian Alps, considering both climate change and land cover transformations. Simulations were done with the Dynamic Revised Universal Soil Loss Equation (D-RUSLE) model, which is able to account also for snow cover and land cover dynamics simulated with automatic machine learning. Results confirm that land cover projections, usually ignored in these studies, might have a significant impact on the estimates of future soil erosion. Our scenario analysis for 2100 shows that if the mean annual precipitation does not change significantly and temperature increases no more than 1.5-2.0 °C, then the erosion rate will decrease by 67% for about half of the study area. At the other extreme, if the mean annual precipitation increases by more than 8% and the temperature increases by more than 4.0 °C, then about three-quarters of the study area increases the erosion rate by 92%. What clearly emerges from the study is that regions with higher erosion anomalies (positive and negative) are expected to expand in the future, and their patterns will be modulated by future land transformations.
... These models are many types that include static or dynamic, spatial or non-spatial, inductive or deductive, agent-based or pattern-based models (Gaucherel and Houet, 2009). LUCC models in general have three sub-models for changing land demand, transition potential, and allocation of change (Eastman et al., 2005). Generation of transition potentials can be modeled using a logistic regression (LR) as used in the CLUE-S (Conversion of Land Use and Its Effects at Small Regional Extent) model (Verburg et al., 1999) and Land Change Modeler (LCM) (Eastman, 2006). ...
... Other methods use empirically derived probabilities as used in the GEOMOD model (Pontius et al., 2001); Weights of evidence as used in DINAMICA (Soares-Filhoetal., 2002), and Multi-Layer Perceptron (MLP) as used in the LCM (Eastman, 2006). Two procedures, MLP and LR, are viable techniques, with MLP being more robust than LR (Eastman et al., 2005). Also, MLP can model non-linear relationships between explanatory variables (Eastman, 2009). ...
Conversion of forest land to farmland in the Hyrcanian forest of northern Iran increases the nutrient input, especially the phosphorus (P) nutrient, thus impacting the water quality. Modeling the effects of forest loss on surface water quality provides valuable information for forest management. This study predicts the future impacts of forest loss between 2010 and 2040 on P loading in the Tajan River watershed at the sub-watershed level. To understand drivers of the land cover, we used Land Change Modeler (LCM) combining with the Soil Water Assessment Tool (SWAT) model to simulate the impacts of land use change on P loading. We characterized priority management areas for locating comprehensive and cost-effective management practices at the sub-watershed level. Results show that agricultural expansion has led to an intense deforestation. During the future period 2010-2040, forest area is expected to decrease by 34,739 hm 2. And the areas of pasture and agriculture are expected to increase by 7668 and 27,071 hm 2 , respectively. In most sub-watersheds, P pollution will be intensified with the increase in deforestation by the year 2040. And the P concentration is expected to increase from 0.08 to 2.30 mg/L in all of sub-watersheds by the year 2040. It should be noted that the phosphorous concentration exceeds the American Public Health Association's (APHA) water quality standard of 0.2 mg/L for P in drinking water in both current and future scenarios in the Tajan River watershed. Only 30% of sub-watersheds will comply with the water quality standards by the year 2040. The finding of the present study highlights the importance of conserving forest area to maintain a stable water quality. Citation: Fatemeh RAJAEI, Abbas E SARI, Abdolrassoul SALMANMAHINY, Timothy O RANDHIR, Majid DELAVAR, Reza D BEHROOZ, Alireza M BAVANI. 2018. Simulating long-term effect of Hyrcanian forest loss on phosphorus loading at the sub-watershed level. Journal of Arid Land, https://doi.
... These models are many types that include static or dynamic, spatial or non-spatial, inductive or deductive, agent-based or pattern-based models (Gaucherel and Houet, 2009). LUCC models in general have three sub-models for changing land demand, transition potential, and allocation of change (Eastman et al., 2005). Generation of transition potentials can be modeled using a logistic regression (LR) as used in the CLUE-S (Conversion of Land Use and Its Effects at Small Regional Extent) model (Verburg et al., 1999) and Land Change Modeler (LCM) (Eastman, 2006). ...
... Other methods use empirically derived probabilities as used in the GEOMOD model (Pontius et al., 2001); Weights of evidence as used in DINAMICA (Soares-Filhoetal., 2002), and Multi-Layer Perceptron (MLP) as used in the LCM (Eastman, 2006). Two procedures, MLP and LR, are viable techniques, with MLP being more robust than LR (Eastman et al., 2005). Also, MLP can model non-linear relationships between explanatory variables (Eastman, 2009). ...
... El procedimiento de modelación se dividió en cuatro pasos: 1) clasificación de imágenes satelitales para la obtención de insumos de cobertura y uso del suelo; 2) análisis de la intensidad de cambio de cobertura y uso del suelo; 3) predicción del cambio de cobertura y uso del suelo; y 4) proceso de validación de mapas de cobertura. De igual modo, se usó el software científico TerrSet, en específico la herramienta Land Change Modeler (lcm), un modelo enfocado en sentido primordial a los problemas del cambio de cobertura y uso del suelo, al igual que a las necesidades de análisis de conservación de la biodiversidad [Eastman et al., 2005]. El modelo sirve para elaborar una proyección espacialmente explícita hacia el futuro, como posibles hechos, es decir, futuros inciertos, pero útiles ante todo para la toma de decisiones y su evaluación [Sharp et al., 2016]. ...
... There are several methods for running transition potential modeling (Aburas et al., 2019;Brown et al., 2012). Among these techniques, the Multi-Layer Perceptron Artificial Neural Network (MLP-ANN) is the most robust and practical (Eastman et al., 2005). ...
Human activities are prone to be the main drivers of land use land cover (LULC) changes, which have cascading effects on the environment and ecosystem services. The main objective of this study is to assess the historical spatiotemporal distributions of LULC changes as well as estimated future scenarios for 2035 and 2045 by considering the explanatory variables of LULC changes in Zanjan province, Iran. The LULC time-series technique was applied using three Landsat images for the years 1987, 2002, and 2019. Multi-layer Perceptron Artificial Neural Network (MLP-ANN) is applied to model the relationships between LULC transitions and explanatory variables. Future land demand was calculated using a Markov chain matrix and multi-objective land optimization in a hybrid simulation model. Validation of the model’s outcome was performed using the Figure of Merit index. The residential area in 1987 was 6406.02 ha which increased to 22,857.48 ha in 2019 with an average growth rate of 3.97%. Agriculture increased annually by 1.24% and expanded to 149% (890,433 ha) of the area occupied in 1987. Rangeland showed a decline concerning its area, with only about 77% (1,502,201 ha) of its area in 1987 (1,166,767 ha) remaining in 2019. Between 1987 and 2019, the significant net change was a conversion from rangeland to agricultural areas (298,511 ha). Water bodies were 8 ha in 1987, which increased to 1363 ha in 2019, with an annual growth rate of 15.9%. The projected LULC map shows the rangeland will further degrade from 52.43% in 2019 to 48.75% in 2045, while agricultural land and residential areas would be expanded to 940,754 ha and 34,727 ha in 2045 from 890,434 ha and 22,887 ha in 2019. The findings of this study provide useful information for the development of an effective plan for the study area.
... In the ecological protection scenario (SP), the government will focus on ecological management and land restoration. We used Cellular Automata-Markov (CA-Markov) model (Eastman et al., 2005) to simulate the LULC structure in 2030 under SN, and build a multi-objective model to calculate the LULC structure under optimization scenarios. ...
... In predicting future land-use change, the sub-modelling of transitions has been practiced in LULCC studies at various spatiotemporal scales. The generality of LULCC models embraces a vicinity within which a trial is created for modelling the probabilities of land-use transitions from one category to another (Eastman et al. 2005). In the study, the model employed the MLP CA-Markov Chain algorithm multiple times for the evaluation of earlier (2006) and later (2012) LULCs, calibration period (2018), and driver variables. ...
Land use and land cover change (LULCC) is considered one of the major drivers of climate change, although climate change can also foster direct or indirect influences leading to LULCC. The objective of the presented study is to offer a strategic observation frame as the land use and land cover (LULC) transitions are grouped to define the cover flows (CFs). The Küçük Menderes River Basin (KMRB), which is located in the west of Turkey was examined as the case study. Through LULCC modelling via the employment of multi-layer perceptron (MLP), cellular automata (CA), and Markov Chain methods, future LULC maps were projected up to the horizon of 2050. Hydrologic responses of the basin to LULCC were determined by the developed hydrologic model, which is generated by the Soil and Water Assessment Tool (SWAT). The superimposed impacts of the examined effects of LULCC have been investigated by the CF types. This way, the individual impacts of the CFs have been assessed. In the case of the KMRB, projected annual runoffs for the year 2050 cover map represent a 9.06% reduction and the major responsible CF type for this reduction is the conversion from forest to non-irrigated agricultural land cover by 22.90%.
HIGHLIGHTS
Future water availability of the KMRB is investigated by the defined cover flows.;
About a 9% reduction in surface water volume is expected by the year 2050 at the KMRB outlet.;
Conversion to non-irrigated agriculture is the major cause of the water volume reduction in the basin.;
... A low Cramer's value shows that a potential driver variable is not suiTable The P-value shows the likelihood that Cramer's value is not nil. Therefore, a low P-value is not a good indication of variable value, but it can be rejected by a high value (Eastman, 2006;Eastman et al. 2005). ...
World’s oldest Aravalli range provides various ecosystem services that are continually vulnerable due to human-induced intervention. Urgently, it needs to be addressed the study of ecologically sensitive areas change. Therefore, the current study assessed a long-time-series land-use dynamics pattern between 1975 and 2059 using remote sensing and machine learning-based approaches. The land-use trend was assessed using the CART (Classification and Regression Tree) supervised classification technique on the Google Earth Engine (GEE) platform for the last 44 years. The MLP-NN (Multilayer Perceptron-Neural Network) algorithm has simulated the 2019 land use map and the upcoming 40-years of decadal land use prediction using the CA (Cellular Automata) Markov model in the LCM (Land Change Modeler) TerrSet platform. We highlight the conversion between spatial land cover and land use patterns and their conversion between classes. The results show that 3676 km² and 776.8 km² (i.e., 4.86% and 1.02%) converted into barren land and settlement from 1975 to 2019, 5772.7 km² (7.63%) of forest land has decreased in Aravalli. In 2059, a total of 16360.8 km² (21.64%) of forest land will be converted to a settlement class. Like mining and settlement, these human interventions are induced tacitly, which provokes ecological imbalances by breaching environmental integrity and hampering the progress of Sustainable Development Goals. This study would help the planers of the cities, forest managers, and the government develop the conservation management plan and sustainable city expansion projects.
... Transition potential modeling is used to find suitable explanatory variables and generate transition potential maps based on the calculation of these variables [34]. The preliminarily selected variables are shown in Table 6, including evidence likelihood-normalized past changes, topographic factors (elevation and slope degree), the distance to each category, the distance from roads, and the distribution density of each category. ...
Protecting areas of important ecological value is one of the main approaches to safeguarding the Earth’s ecosystems. However, the long-term effectiveness of protected areas is often uncertain. Focusing on China’s ecological conservation redline policy (Eco-redline policy) introduced in recent years, this study attempted to examine the effectiveness of alternative policy interventions and their implications on future land-use and land-cover (LULC) patterns. A scenario analysis was employed to elucidate the implications of different policy interventions for Chongqing capital, one of the most representative cities in China. These interventions considered the spatial extent of Eco-redline areas (ERAs) and the management intensity within these areas. LULC data for two different periods from 2000 (first year) to 2010 (end year) were derived from satellite images and then used for future (2050) LULC projections, incorporating the various policy interventions. Furthermore, several landscape indices, including the shape complexity, contrast, and aggregation of forest patches were calculated for each scenario. After comparing the scenarios, our analysis suggests that the current extent of ERAs may not be sufficient, although their management intensity is. Therefore, we suggest that during the optimization of the Eco-redline policy, ERAs are gradually increased while maintaining their current management intensity.
... Thus the ROC analysis is useful for cases in which the scientist wants to see how well the suitability map portrays the location of a particular category but does not have an estimate of the quantity of the category. Eastman (2005) used the ROC to test the performance of the land use transition potential of twelve GIS-based models for predicting the future state of an area. ...
The study characterized urban expansion in Lokoja; simulated the pattern of future urban expansion for year 2021 and used a GIS-based multi-criteria evaluation to derive a suitable pattern for sustainable urban expansion in the area. These formed the basis for performing a pairwise spatial comparisons among the derived land suitability map, the existing land use map, the simulated land use map and the existing land use plan of Lokoja. The foregoing were done with the overall aim of identifying an optimal spatial pattern for sustainable urban expansion in the Lokoja.
Primary and secondary data were used for this study. The primary data was obtained using a Global Positioning System (GPS) receiver. Secondary data included Landsat ETM+ imageries (of years 1990, 2001, and 2011), the Advanced Spaceborne Thermal Emissions and Reflections Radiometer’s Global Digital Elevation Model (ASTER GDEM) imagery, the Lokoja land use plan. The secondary datasets were obtained from the appropriate authorities. The satellite imageries were used to characterize urban expansion in the study area, simulate the pattern of future urban expansion, and derive relevant land suitability criteria. The Lokoja land use plan and neighbourhood map were assessed in terms of their spatial agreement with the land suitability map derived from this study. Data collected were analyzed using appropriate geospatial techniques. This study used the Weighted Linear Combination (WLC) decision rule in aggregating the various spatial criteria into a final urban expansion suitability index. Linear Spectral Mixture Analysis (LSMA) aided in characterizing urban expansion while Cellular Automata (CA) was used to simulate the pattern of future urban expansion in Lokoja. The Relative Operating Characteristic (ROC) was used in the pairwise quantification of spatial agreement.
The study revealed that the area of urban areas increased from 2.79% in year 1990 to 3.40% in year 2001 and 5.03% in year 2011. Based on these, the future spatial extent of urban land uses in the study area for year 2021 was simulated to be 5.85%. Of the urban land uses, Medium Density Residential area occupied the largest area in all the study epochs but year 2011 where Low Density Residential area occupied the largest area (i.e. 1.79% of the study area). Despite this, Medium Density Residential area was simulated to possess the largest spatial extent in year 2021. 3.18% of the study area experienced major changes in urban land uses across the three study epochs, while only 0.64% did not experience any urban land use change at all. On integrating all the 11land suitability criteria, the highest suitability score obtained was 76%, with 56.58% of the study area being suitable (with scores of above 55%) for sustainable urban expansion. 25.71% (107.45sqKM) of the study area was highly suitable with suitability values of 64% – 76% while 7.50% (31.33 sqKM) was marginally suitable with suitability scores of 29% – 44%. Finally, the derived land suitability map of this study possessed an agreement of 51%, 64% and 66% with the current land use plan, the current urban land use map and the simulated urban land use of year 2021 respectively.
The study concluded that there is a sufficient spatial agreement between the current urban land uses and the derived land suitability areas of this study. Interestingly, if current trends in urban expansion continue, this degree of spatial agreement is expected to increase in year 2021. Two of the designated land uses possessed a very poor degree of spatial agreement with the derived suitability map of this study.
Keywords: Sub-pixel Land Use/Cover Classification, Spectral Mixture Analysis, Cellular Automata, Multicriteria Evaluation, Sustainable Urban Expansion
... The state of a cell or a delimited geographic area primarily depends on its initial state and subsequently of the state of the neighbouring cells. It is possible to do this kind of simulation with different GIS, the most well-known is the software IDRISI (see Eastman et al., 2005;Eastman et al, 2008;Eastman 2009). ...
This doctoral dissertation is an interdisciplinary research work at the intersection of geography and innovation studies through the prism of two subjects: diffusion of sustainable innovations and urban resilience to the energy transition. The research work was developed in the Swiss Alps and in the South Region of France in order to deploy comparative analyses. This Ph.D dissertation investigates the regions receptivity to sustainable innovations, specifically renewable energy technologies (RET) such as solar photovoltaics, solar thermal collectors in the Swiss Alps and electric and hybrid vehicles in Switzerland at the national level. As well as the research was developed in the South Region of France were six innovation indicators were analysed: solar photovoltaics, solar thermal collectors, wind power, small and big hydroelectric power plants, biogas and biomass. The research aimed at improving our understanding on regions’ ability to integrate these innovations into their dynamics and the embedded urban resilience to the energy transition, to adapt to change, accommodating disruptions in the diffusion process and develop new spatial diffusion paths. The research questions aimed at underpinning our understanding on the network effects of the RET diffusion and on the potential insights that spatial information might provide regarding such diffusion processes. The underlying assumptions were that RET diffuses across scales in a non-random fashion and describe a preferential attachment mechanism. The implication of the latter assumption is that the urban renewable energy systems exhibit fractality, which is the signature of self-organized systems. By definition, resilient systems are self-organized, so in this context, the innovation systems are analyzed in the framework of urban resilience to the energy transition. Thus, a further assumption is made and proposes that more innovative places and locations are more resilient than less innovative locations. The methodological approaches to address these research questions and verify the assumptions are described as follows. In the Swiss region a model called ‘Spatial Preferential Attachment’ (SPA) was created based on spatial interaction theory, relying on a gravity model that was built through agent-based modelling and systems dynamics approaches. The integration of the gravity model with the spatial information of the RET allowed to build a spatial network, which simulated the urban energy system of the region. The results allowed to accept the assumptions in which a preferential mechanism in the diffusion process take place, since the spatial diffusion distribution follow power laws. The model was also applied in Switzerland and in the South Region of France, obtaining similar results, following multi-level and hierarchical mechanisms. These results are in line with the path development theory proposed by economic geographers, where the specific-place legacy has at least a partial impact in the future intensity of diffusion processesThese results are important within the sustainability paradigm from a research perspective and challenging for the current innovation framework, the so-called Transformative Change, which aims at establishing a fairer view on socio-economic and environmental issues. The preferential attachment mechanisms in RET diffusion imply that there are hubs of innovation ruled by urban scaling laws that put in disadvantage other locations. The SPA was also used to simulate the urban resilience to the energy transition in the Swiss canton of Valais. The energy urban network was ‘attacked’ by removing the hubs of the structure and the simulations showed that the system could reorganized itself at global level, showing strong sings of resilience however not at local level. Resilient systems are self-organized however it does not imply that resilient itself is fractal as differences were found from a multiscale spatial perspective.
... The state of a cell or a delimited geographic area primarily depends on its initial state and subsequently of the state of the neighbouring cells. It is possible to do this kind of simulation with different GIS, the most well-known is the software IDRISI (see Eastman et al., 2005;Eastman et al, 2008;Eastman 2009). ...
This doctoral dissertation is an interdisciplinary research work at the intersection of geography and innovation studies through the prism of two subjects: diffusion of sustainable innovations and urban resilience to the energy transition. The research work was developed in the Swiss Alps and in the South Region of France in order to deploy comparative analyses. This Ph.D. dissertation investigates the region's receptivity to sustainable innovations, specifically renewable energy technologies (RET) such as solar photovoltaics, solar thermal collectors in the Swiss Alps, and electric and hybrid vehicles in Switzerland at the national level. As well as the research was developed in the South Region of France where six innovation indicators were analysed: solar photovoltaics, solar thermal collectors, wind power, small and big hydroelectric power plants, biogas, and biomass. The research aimed at improving our understanding of regions’ ability to integrate these innovations into their dynamics and the embedded urban resilience to the energy transition, to adapt to change, accommodating disruptions in the diffusion process and develop new spatial diffusion paths.
The research questions aimed at underpinning our understanding on the network effects of the RET diffusion and on the potential insights that spatial information might provide regarding such diffusion processes. The underlying assumptions were that RET diffuses across scales in a non-random fashion and describes a preferential attachment mechanism. The implication of the latter assumption is that the urban renewable energy systems exhibit fractality, which is the signature of self-organized systems. By definition, resilient systems are self-organized, so in this context, the innovation systems are analyzed in the framework of urban resilience to the energy transition. Thus, a further assumption is made and proposes that more innovative places and locations are more resilient than less innovative locations.
The methodological approaches to address these research questions and verify the assumptions are described as follows. In the Swiss region a model called ‘Spatial Preferential Attachment’ (SPA) was created based on spatial interaction theory, relying on a gravity model that was built through agent-based modelling and systems dynamics approaches. The integration of the gravity model with the spatial information of the RET allowed building a spatial network, which simulated the urban energy system of the region. The results allowed us to accept the assumptions in which a preferential mechanism in the diffusion process takes place since the spatial diffusion distribution follows power laws. The model was also applied in Switzerland and in the South Region of France, obtaining similar results, following multi-level and hierarchical mechanisms. These results are in line with the path development theory proposed by economic geographers, where the specific-place legacy has at least a partial impact on the future intensity of diffusion processes.
These results are important within the sustainability paradigm from a research perspective and challenging for the current innovation framework, the so-called Transformative Change, which aims at establishing a fairer view on socio-economic and environmental issues. The preferential attachment mechanisms in RET diffusion imply that there are hubs of innovation ruled by urban scaling laws that put in disadvantage other locations. The SPA was also used to simulate the urban resilience to energy transition in the Swiss canton of Valais. The energy urban network was ‘attacked’ by removing the hubs of the structure and the simulations showed that the system could reorganize itself at a global level, showing strong signs of resilience however not at the local level. Resilient systems are self-organized however it does not imply that resilient itself is fractal as differences were found from a multiscale spatial perspective.
... The TPR describes how good the model is at predicting the positive class when the actual outcome is positive, while the FPR is the proportion of the units with a known negative condition for which the predicted condition is positive. The ROC statistic metric is the AUC (Area Under the Curve) values that range from 0.5 (random) to 1.0 (perfect predictive power) (Eastman et al. 2005). For the quantitative validation, a simple pixel-based comparison between the real data (validation dataset) and predicted probability was analysed. ...
The paper investigates built-up areas expansion after the 1990 in one of the highly urbanized regions of Romania - Romanian Plain, in order to explore the urban sprawl phenomena and its temporal and regional disparities in relation to some of the main distance driving factors. The research uses Landsat 4/5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper (ETM), and Landsat 8 Operational Land Imager (OLI) imagery to derive built-up areas and quantify their expansion over time in relation to fourteen distance explanatory factors: i.e., previous built-up areas, main road infrastructure, Bucharest city’s boundary, location of the urban centres classified according to demographic size and main economic function, forest land and water bodies. To estimate the influence of the predictors, the binary logistic regression was applied. Furthermore, to estimate the effectiveness of the predictor set in the variation of built-up areas expansion, the pseudo R² was calculated and discussed. Moreover, to understand the future potential trend of urban sprawl and its spatial pattern, the probability maps were generated by integrating the regression coefficients of the statistically significant predictors into the spatial modeling. For the results performance assessment, the statistic Receiver Operating Characteristic and the pixel-based comparison between the real and predicted data were used. To assess possible differences at spatial and temporal scale, the analysis was carried out at regional level, for two periods: 1990–2002 and 2002–2018. In general, our findings show inverse relationship between the distance driving factors and built-up areas expansion, but the estimated predictive power suggests important disparities within the study area over the analysed periods. Overall, the statistical analysis indicate that the distance to previous build-up areas, distance to road infrastructure, distance to Bucharest and other large urban centres, and distance to urban centres with dominant industrial and service functions were more influential to urban sprawl after 1990. Furthermore, the predicted spatial data shows the highest potential of urban sprawl in the future around Bucharest, in the proximity of existing built-up areas and road infrastructure. Because of its predictive character, the present study is to be a useful tool for land managers and policy makers.
... 2.5. Land use/cover change analysis A change matrix was estimated for each of the three analyzed periods (1986-1996/ 1996-2005/2005-2015), cross-tabulation procedure between classifications was processed and area change, gains, losses and persistence were calculated for each period with Land Change Modeler IDRISI Extension (Eastman, Solorzano, and Van Fossen 2005) on IDRISI Selva software (Clark Labs 2012). Growth and decrease rates were estimated for each category and period. ...
Urbanization and agricultural land expansion are the largest drivers of global land
cover change. Here, we aimed to quantify three decades of land-use/land-cover
change across one of the main horticultural regions of South America. We assessed
landscape change implementing a supervised classification workflow on Landsat
satellite imagery (1986, 1996, 2005 and 2015). Between 1986 and 2015,
horticulture extent decreased (51.47%) at the expense of a high increase in
greenhouses (2652.83%). Additionally, high density urbanization experienced a
strong expansion (111.58%), while low density urbanization increased only
between 1986 and 2005, replacing natural grassland, herbaceous parks and
livestock. These results demonstrate a regional urban growth and productive
intensification process that echoes similar global processes with consequential
losses of open field horticultural areas and a non-equitable distribution of seminatural
areas in this region. Adequate territorial planning toward ecological resilient
territories that consider ecological processes and prioritize semi-natural vegetation
cover is urgently needed.
... Multitemporal aspects (time series) are the pivot of analysis of change because they incorporate initial or before (t1) and after (t2) conditions. Changes are defined from the transition of the distribution of dominant objects between t1 and t2 [41]. t1 in this study is 2015 and t2 is 2019. ...
The growth of human occupations in coastal areas and climate change impact have changed the dynamics of seagrass cover and accelerated the damage to coral reefs globally. For these reasons, coastal management measures need to be developed and renewed to preserve the state of seagrass beds and coral reefs. An example includes the improvement of spatial and multitemporal analyses. This study sought to analyze changes in seagrass cover and damages to coral reefs in Gili Sumber Kima, Buleleng Regency, Bali based on multitemporal Sentinel 2A-MSI imagery. The algorithms of a machine learning, Random Forest (RF), and a Support Vector Machine (SVM) were used to classify the benthic habitats (seagrass beds and coral reefs). Also, a change detection analysis was performed to identify the pattern and the extent to which seagrass beds had changed. The multispectral classification of, particularly, coral reefs was used to explain the condition of this benthic habitat. The results showed +-70% to +-83% accuracies of estimated seagrass cover, and the change detection analysis revealed three directions of change, namely an increase of 27.9 ha, a decrease by 86 ha, and a preserved state in 157 ha of seagrass cover. The product of coral reefs mapping had an accuracy of 42%, and the coral reefs in Gili Sumber Kima were split almost equally between the good (1505 ha) and damaged ones (1397 ha). With the spatial information on seagrass beds and coral reefs in every region, the ecological functions of the coast can be assessed more straightforwardly and appropriately incorporated as the basis for monitoring the dynamics of resources and coastal area management.
... In subsequent research (Joorabian Shooshtari et al., 2018), a similarity-weighted instance-based machine learning algorithm in LCM was successfully applied to predict multiple transitions among land cover classes. In the current work, MLP neural network in LCM, known to be more robust than other algorithms, was used to generate a transition potential surface for each land cover transition (Eastman et al., 2005;Sangermano et al., 2012). Validation of the MLP after 10,000 iterations showed the accuracy rate to be above 88% in all submodels, and the lowest and highest values were 88.10 and 95.78%, respectively. ...
The spatial and temporal dimensions of environmental impacts of climate and land cover changes are two significant factors altering hydrological processes. Studying the effects of these factors on water quality, provides important insight for water resource management and optimizing land planning given increasing water scarcity and water pollution. The impact of land cover and climate changes on surface water quality was assessed for the Neka River basin in Northern Iran. The widely used Soil and Water Assessment Tool (SWAT) was applied for pollutant modeling, and was calibrated using the Sequential Uncertainty Fitting (SUFI-2) algorithm. An ensemble of 17 CMIP5 climate models under two IPCC greenhouse gas emission scenarios were selected, and future land cover change (LCC) was modeled based on the evolution that occurred in the last decades. We simulated the impacts of climate change (CC) and LCC on sediment, nitrate, and phosphate for the 2035–2065 time slice. The annual loads of sediment, phosphate, and nitrate are projected to decrease under both CC scenarios based on the inter-model average, and generally follow a pattern similar to the change in river discharge. Nitrate concentrations show an increase across all seasons, while the sediment and phosphate concentrations increase in winter and autumn under CC conditions. Results indicate that pollutants are expected to increase under LCC alone, mainly due to the expansion of the cultivated areas. Overall, it seems CC has a greater impact than LCC on the variation of water quality variables in the Neka River basin. With a combined change in climate and land cover, the annual nitrate concentrations are expected to increase by + 19.7% and + 17.9%, under RCP 4.5 and RCP 8.5, respectively. The combined impacts of the CC and LCC caused a decline in the annual sediment and phosphate concentrations by −10.1% and −2.2% under RCP 4.5 and −9%, and −3.2% under RCP 8.5, respectively.
... The spatio-temporal change analysis is a complex and nonlinear practice with the specific direct and indirect drivers at various scales that are very diverse (Lambin et al., 2003;Kolb et al., 2013). The competency of a model and its success rate depends on the studying or involvement of driving factors, cross-scale dynamics, various levels and sub-levels of analysis, spatial relations and neighbourhood effects, integration level, and temporal changing direction of landscapes and its drivers (Lambin et al., 2003;Verburg et al., 2004;Bürgiet al., 2004;Eastman et al., 2005;Kolb et al., 2013). The complexity, inherent dynamic characteristics, and ambiguity of natural systems require a conceptualized demonstration of sustainable LULC practices based on modelling processes in rapidly growing regions. ...
Assessing the performance of land change simulation models is a critical step when predicting the future landscape scenario. The study was conducted in the district of Varanasi, Uttar Pradesh, India because the city being "the oldest living city in the world" attracts a vast population to reside here for short and long-term, leaving the city's ecosystem more exposed to fragility and less resilient. In this work, an approach based on landscape metrics is introduced for comparing the performance of the ensemble models designed to simulate the landscape changes. A set of landscape metrics were applied in this study that offered comprehensive information on the performance of scenario-based simulation models from the viewpoint of the spatial ordering of simulated results against the related reference maps. A supervised support vector machine classification technique was applied to derive the LULC maps using Landsat satellite images of the year 1988, 2001, and 2015. The LULC maps of 1988 and 2001 were used to simulate the LULC scenario for 2015 using three Markov chain-based simulation models namely, multi-layer perceptron-Markov chain (MLP_Markov), cellular automata-Markov chain (CA_Markov), and stochastic-Markov chain (ST_Markov) respectively. The mean relative error (MRE), as a measure of the success of simulation models, was calculated for metrics. The MRE values at both the class and landscape levels were accounted for 21.63 and 11.45% respectively using MLP_Markov simulation model. The MRE values at both the class and landscape levels were accounted for 39.61 and 28.31% respectively using CA_Markov simulation model. The MRE values at both the class and landscape levels were accounted for 55.36 and 45.75% respectively using ST_Markov simulation model. The MRE values considered at class and landscape levels are further evaluated qualitatively for comparing the performance of simulation models. The results indicate that the MLP_Markov performed excellently, followed by CA_Markov and ST_Markov simulation models. This work showed an ordered and multi-level spatial evaluation of the models' performance into the decision-making process of selecting the optimum approach among them. Landscape metrics as a vital characteristic of the utilized method, employ the maximum potential of the reference and simulated layers for a performance evaluation process. It extends the insight into the main strengths and drawbacks of a specific model when simulating the spatio-temporal pattern. The quantified information of transition among landscape categories also provides land policy managers a better perception to build a sustainable city master plan.
... We modeled each transition using the multilayer perceptron (MLP) method in Idrisi [18]. After running each sub-model, we obtained seven transition maps and their corresponding accuracy values. ...
The aims of this study were to identify the dynamics of land use change, the factors associated to these changes, and potential transformations of paramo and Andean forest, through the modeling of land use change scenarios in the department of Boyacá, Colombia. Throughout the classification of satellite images, we assessed land use change in two time periods: 1998 to 2010 and 2010 to 2018. Seven transition sub-models were analyzed and associated to 36 explanatory variables. Three future scenarios of land use change were projected for the years 2030 and 2050: trend, agricultural expansion and conservation scenarios. We found a gradual reduction in paramos and Andean forests, together with an increase in secondary vegetation. The most relevant variables explaining land use change were: elevation, distance to roads and distance to protected areas. The scenario with the greatest impact in paramos and Andean forest was Agricultural Expansion, where forest would have a loss of 29% and 41% for 2030 and 2050, and Paramos 44% and 59% for the same years. Forest and paramos in the central eastern area showed critical losses and highly fragmented distributions in all scenarios; hence, we recommend focusing conservation efforts in these areas.
... Desde un punto de vista más general del presente estudio, se puede decir que, en cuanto a las cuatro técnicas que presenta MOLUSCE, el modelo obtenido con redes neuronales -MLP combinado con autómatas celulares (AC -MLP) fue el que mejor resultado alcanzó, motivo por el cual fue escogido para compararlo con los otros modelamientos. El hecho de que AC -MLP haya sido el que ajustó de mejor forma el crecimiento urbano de Macas (en MOLUSCE), concuerda con lo hallado en los resultados de LCM, ya que de sus tres métodos: LR, MLP y SimWeight, las redes neuronales con MLP alcanzaron el segundo mejor rendimiento, conclusión similar a la que llegó Eastman et al., (2005), quienes compararon varias técnicas (pesos de evidencia, probabilidades empíricas, regresión logística y redes neuronales -MLP), y determinaron que MLP modeló de mejor forma el potencial de transición. Sin embargo, para este caso de estudio, SimWeight obtuvo un mejor rendimiento que MLP por las determinaciones mencionadas en el párrafo anterior, y en consecuencia, fue seleccionado como el modelo óptimo para modelar el crecimiento tendencial urbano de Macas. ...
El crecimiento de las ciudades es uno de los principales problemas con gran impacto sobre el territorio, cambiando el uso y cobertura del suelo, y dando lugar a nuevas afectaciones ambientales y sociales. La ciudad de Macas (Ecuador), ha tenido un crecimiento poco regulado y disperso, por lo que predecir su expansión coadyuvaría a preveer posibles inconvenientes a futuro. El objetivo del estudio fue, generar un modelo de crecimiento tendencial urbano de Macas al año 2030, mediante modelación espacial multivariable, para que sirva como una herramienta en la delimitación del suelo rural de expansión urbana según la LOOTUGS del Ecuador. Se aplicaron diversas técnicas de simulación espacial, de las cuales tres fueron las escogidas para su comparación y selección del mejor modelo, como fueron: autómatas celulares con perceptrón multicapa (AC – MLP) y con algoritmos genéticos (AC – AG), y aprendizaje ponderado por similitud (SimWeight). Los resultados de la validación mostraron un 31%, 47% y 54% de similitud para AC – AG, AC – MLP y SimWeight, respectivamente, por tanto, éste último fue el mejor modelo. Con la mancha urbana resultante, se identificó una tendencia de expansión hacia el Norte, y delimitó una propuesta del suelo de expansión urbana para Macas, la misma que está adyacente al área urbana actual. El modelo de ciudad deseado, las políticas que se tomen y la inclusión de este tipo de análisis espacial, permitirán planificar de una forma sostenible y ordenada a la ciudad.
... The AUC values range from 0.5 to 1.0: a measure with perfect predictive power would yield a value of 1.0, while one with no power (random) would yield a value of 0.5 [17]. ...
Forest-cover dynamics is of wide concern due to its role in climate change, biodiversity losses, water balance and land degradation, as well as social and economic development. Hence, exploring land-use/cover dynamic is important in order to improve our understanding of the causes of forest-cover change and to detect the future trend. Furthermore, projecting a future land-use/cover pattern can help identifying potential areas where forest-cover change will occur in the future and the potential consequences of these processes in order to improve land-use planning and policies. Similar to other East European countries, Romania is experiencing rapid land-use/cover changes after the breakdown of socialism; a clear trend was registered by deforestation, which reflects the consequences of a continuous forests dynamics and little environmental care. Consequently, this study, carried out in order to analyse the potential future cover-change, resulted in the land-use/cover scenario (2007–2050) simulated using CLUE-S (the Conversion of Land Use and its Effects at Small regional extent) modelling framework, applied to development regions in Romania. Overall, the model results in different spatial patterns of land-use/cover change, projecting a slight increase in the forest-cover area of about 82,000 ha. Furthermore, the model simulated widespread deforestation, mainly in relation to agricultural land expansion. The area under the curve (AUC) for the relative operating characteristic (ROC) and the Kappa simulation (KSimulation) were used to assess the predictive power of the determinant factors included and to evaluate the spatial performance of the model. The obtained ROC/AUC values (0.83–0.88) indicate the great power of the determinant factors to explain the forest-cover pattern in the area. Furthermore, the KSimulation scores (0.69–0.79) highlight the potential of the CLUE-S model to simulate future forest-cover change in relation to the other land-use/cover categories. The results can provide useful inputs for effective forest resource management and environmental policies. Moreover, the spatial data obtained can contribute to exploring future potential environmental implications (e.g. assessing landslide and flood hazard scenarios, forest biomass dynamics and their impact on carbon allocation, or the impact of forest-cover change on ecosystem services).
... O LCM permite a análise, previsão e validação dos mapas de alteração de uso de solo. Usa como dados de entrada os mapas de dois períodos de tempo distintos contendo as mesmas classes de uso de solo, as mesmas legendas, são localizados nas mesmas áreas de estudo, e têm as mesmas dimensões espaciais, ou seja, a mesma resolução e sistema de coordenadas (Eastman et al., 2005). Neste estudo, os dados de entrada são os mapas CLC de 2000 e 2012, contendo oito categorias: 1 -Urbano contínuo, 2 -Urbano descontínuo, 3 -Edificado não-urbano (industrial, comercial, vias de transporte, minas, aterros e estaleiros de obras), 4 -Agricultura, 5 -Floresta e áreas semi-naturais, 6 -Zonas húmidas, 7 -Corpos de água e 8 -Vegetação artificial ou não cultivada, pelo que foi necessário reagrupar as 44 classes presentes na CLC (nível III). ...
Being security assumed as a basic right of citizens in the current model of democratic rule of law, optimal resources allocation altogether with budgetary constraints are a key component. In fact optimal resources allocation and budgetary constraints oblige an increasingly careful strategic management, adapted to demographic reality. The SIM4SECURITY project aims to build a technological solution to support decision making regarding security, based on the development of a GIS model and in the implementation of demographic scenarios. This model will allow policy makers, leaders and forces of command units and services in the planning and rational affectation of resources adjusted to local dynamics in crime prevention and crime fighting. To communicate the SIM4SECURITY results and support decision making, a data visualization and storytelling approach was adopted by creating dashboards containing the various dimensions and perspectives of the information were elaborated and are presented. The obtained outcomes show that dashboards are an important visual tool in the decision-making process by providing meaningful insights regarding security and in the location-allocation of security forces.
... The dynamics and 74 complexity of the ecosystem requires a more complete evaluation of LULCC. The spatial 75 modeling is a technique contemplating alternative scenarios of LULCC, which could contribute to 76 better explain the key processes influencing LULCC ( Pijanowski et al., 2002;Eastman et al., 2005;77 Torrens, 2006;Perez-Vega et al., 2012). Thus, one of the main functions of the LULCC models is 78 the establishment of scenarios, with the aim of changing policies and inadequate practices for the 79 sustainable management of natural resources (DeFries et al., 2007;Berberoğlu et al., 2016). ...
The loss of temperate forests of Mexico has continued in recent decades despite wide recognition of their importance to maintaining biodiversity. This study analyzes land use/land cover change scenarios, using satellite images from the Landsat sensor. Images corresponded to the years 1990, 2005 and 2017. The scenarios were applied for the temperate forests with the aim of getting a better understanding of the patterns in land use/land cover changes. The Support Vector Machine (SVM) multispectral classification technique served to determine the land use/land cover types, which were validated through the Kappa Index. For the simulation of land use/land cover dynamics, a model developed in Dinamica-EGO was used, which uses stochastic models of Markov Chains, Cellular Automata and Weight of Evidences. For the study, a stationary, an optimistic and a pessimistic scenario were proposed. The projections based on the three scenarios were simulated for the year 2050. Five types of land use/land cover were identified and evaluated. They were primary forest, secondary forest, human settlements, areas without vegetation and water bodies. Results from the land use/land cover change analysis show a substantial gain for the secondary forest. The surface area of the primary forest was reduced from 55.8% in 1990 to 37.7% in 2017. Moreover, the three projected scenarios estimate further losses of the surface are for the primary forest, especially under the stationary and pessimistic scenarios. This highlights the importance and probably urgent implementation of conservation and protection measures to preserve these ecosystems and their services. Based on the accuracy obtained and, on the models generated, results from these methodologies can serve as a decision tool to contribute to the sustainable management of the natural resources of a region.
Quantitative analysis of LULC changes and their effects on carbon stock and sequestration is important for mitigating climate change. Therefore, this study examines carbon stock and sequestration in relation to LULC changes using the Land Change Modeler (LCM) and Ecosystem Services Modeler (ESM) in tropical dry deciduous forests of West Bengal, India. The LULC for 2006, 2014, and 2021 were classified using Google Earth Engine (GEE), while LULC changes and predictions were analyzed using LCM. Carbon stock and sequestration for present and future scenarios were estimated using ESM. The highest carbon was stored in forest land (124.167 Mg/ha), and storage outside the forest declined to 13.541 Mg/ha for agricultural land and 0–8.123 Mg/ha for other lands. Carbon stock and economic value decreased from 2006 to 2021, and are likely to decrease further in the future. Forest land is likely to contribute to 94% of future carbon loss in the study region, primarily due to its conversion into agricultural land. The implementation of multiple-species plantations, securing tenure rights, proper management practices, and the strengthening of forest-related policies can enhance carbon stock and sequestration. These spatial-temporal insights will aid in management strategies, and the methodology can be applied to broader contexts.
This research investigates the future dynamics of water yield services in the Gorgan River Basin in the North of Iran by analyzing land cover changes from 1990 to 2020, using Landsat images and predicting up to 2040 with the Land Change Modeler and InVEST model under three scenarios: continuation, conservation, and mitigation. The results indicate significant shifts in agricultural land impacted water yields, which fluctuated from 324.7 million cubic meters (MCM) in 1990 to 279.7 MCM in 2010, before rising to 320.1 MCM by 2020. The study uniquely assesses the effects of land use changes on water yields, projecting a 13.6% increase in water yield by 2040 under the continuation scenario, a 3.9% increase under conservation, and a 1.6% decrease under mitigation, which limits changes on steep slopes to prevent soil erosion and floods. This underscores the interplay between land use, vegetation cover, and water yield, emphasizing strategic land management for water resource preservation and effective watershed management in the GRB.
The aim of this study is to provide a framework for spatial planning aimed at creating resilient cities through land change trends from the past to the future (1990-2020-2050). Additionally, it aims to guide planning strategies by highlighting the current and future status of urban green areas. In this context, the trends for 2050 were evaluated in terms of green area requirements, using the Silifke example. Most of the development areas in Adana-Mersin RegionalPlan are expected to be built. All of these areas consist of agricultural (202 ha) and bare lands (46 ha). It is estimated that by 2050, the populations of Silifke-Taşucu-Kum Neighborhood will be 102,923 and 24,815, respectively. According to spatial planning regulations, the minimum green area within the 2050 urban plan boundaries should be 1,029,230 m² and 248,150 m², respectively. Consequently, decision-makers are expected to determine green area strategies guided by these findings in spatial planning studies.
This study aims at predicting the urban growth of Djelfa, which is the largest Algerian semi-arid city, and assessing its impact on Land Use / Land Cover (LULC). Three satellite datasets (2000, 2010, and 2020) were classified using Maximum Likelihood Classification (MLC). We employed the LULC maps of 2000 and 2010 and integrated four urban growth factors to predict the urban growth of 2020 using Land Change Modeler (LCM) based on Logistic Regression Model (LRM). The predicted urban growth was compared with the observed urban class of 2020 to validate the model. Finally, we predicted future urban growth of 2030, 2040, and 2050. In effect, the urban growth of Djelfa is rapid. Its annual rate was 3.05% from 2000 till 2020 and will be 1.85% between 2020 and 2050. This has caused the loss of 11.88 km², 2.01 km², and 1.76 km² of the steppe, the forest, and the agricultural land respectively, between 2000 and 2020. According to LCM, the steppe, the forest, and the agricultural land will lose 28.33 km², 32.54 km², and 10.84 km² sequentially, between 2020 and 2050.
Urban land use and land cover (LULC) change and climate affect a wide range of environmental health and human wellbeing issues, such as air pollution and soil erosion. However, it is challenging to quantify the effects of LULC change on Particulate Matter (PM) at a metropolitan scale. In this study, LULC maps of Tehran metropolitan region were prepared for the years 2013, 2017 and 2021 using Landsat 8 images. PM10 maps were extracted from an integrated approach of the Local Climate Zone (LCZ) classification scheme and Landsat images using Random Forest algorithm and fuzzy algorithm on the desired dates. The results show extensive changes of PM distributions in the Tehran metropolitan region. The most change occurred in the north of the city and the least change did in the city center. The most growth was LCZ 4 which had grown about 57 times in 8 years. The highest level of PM10 was found in the west and southwest regions while the lowest level was in the north of the city. Also, the analysis of highest average PM10 was in LCZ 10 and the lowest in LCZ G. The analysis of the PM10 average in LCZs and the comparison of PM10 time-series maps showed that the LCZs changes which occurred during the study period have reduced the concentration of air pollution in the Tehran metropolitan region
Land use/land cover change (LULCC) is a major threat that affects the viability of insect populations worldwide yet our estimates of such effects are usually poor. We analysed how LULCC affected the distribution of 49 species of dragonflies and damselflies in the south-central zone in Mexico during the period 2006-2012. For this, we mapped the potential species richness using ecological niche models in order to analyse predicted future changes and determined the effect of LULCC on the current and future habitats of Odonata. We also estimated current incidence of deforestation and projected its effect to 2050 using the Dinamica-EGO program. Having predicted the level of deforestation in the year 2050, we then compared current vs. expected species richness and the factors that determine it. First, roads and urban areas turned out to be the most important drivers of LULCC in our analysis. Second, deterioration occurred at all sites, but species richness remained high despite considerable habitat fragmentation. Third, there is likely to be a high species turnover rate (i.e. a high species richness, but composed of different species) even in areas where there are significant changes in the vegetation. Our work illustrates both a resilience of Odonata to LULCC and provides a useful method for measuring the effects of LULCC on insects.
Research on the uncertainty of Land Use Cover Change (LUCC) models is still limited. Through this paper, we aim to globally characterize the structural uncertainty of four common software packages (CA_Markov, Dinamica EGO, Land Change Modeler, Metronamica) and analyse the options that they offer for uncertainty management. The models have been compared qualitatively, based on their structures and tools, and quantitatively, through a study case for the city of Cape Town. Results proved how each model conceptualised the modelled system in a different way, which led to different outputs. Statistical or automatic approaches did not provide higher repeatability or validation scores than user-driven approaches. The available options for uncertainty management vary depending on the model. Communication of uncertainties is poor across all models.
Context
Evaluation of land cover change (LCC) is commonly done at the pixel level; however, the model’s purpose may be relevant at a different grain size. Thus, the same model may be good for one purpose but inappropriate for another. For conservation applications, it is crucial to assess land change simulations at the grain relevant for the assessment of biodiversity impacts.
Objectives
Our objective is to evaluate land cover change scenarios in Bolivia, at the pixel-level and grain relevant to biodiversity, to inform LCC models for biodiversity assessments.
Methods
We created six deforestation simulations that varied deforestation allocation based on forest management units (national, province, and municipality), ecoregions, and carbon stocks. We evaluated the simulations at the pixel level, and the objective’s relevant grain size through stratified error decomposition. We assessed biodiversity impacts by comparing the quantity of reference and simulated deforestation within species ranges.
Results
The spatial allocation of deforestation differed across simulations; however, their pixel-level error were similar. The province and municipality land change simulations had the lowest allocation errors at the relevant grain despite their large pixel-level errors, and they showed the lowest biodiversity errors. The province simulation provided the best balance identifying both affected species composition and the area of impact.
Conclusions
This work presents evidence of the importance of incorporating information regarding the purpose of the simulation during model evaluation and selection. Error decomposition allowed ignoring irrelevant errors, translating into meaningful assessments of biodiversity impacts. As opposed to pixel-level metrics, stratified errors identified models that characterized biodiversity impacts best.
The loss of temperate forests of Mexico has continued in recent decades despite wide recognition of their importance to maintaining biodiversity. This study analyzes land use/land cover change scenarios, using satellite images from the Landsat sensor. Images corresponded to the years 1990, 2005 and 2017. The scenarios were applied for the temperate forests with the aim of getting a better understanding of the patterns in land use/land cover changes. The Support Vector Machine (SVM) multispectral classification technique served to determine the land use/land cover types, which were validated through the Kappa Index. For the simulation of land use/land cover dynamics, a model developed in Dinamica-EGO was used, which uses stochastic models of Markov Chains, Cellular Automata and Weight of Evidences. For the study, a stationary, an optimistic and a pessimistic scenario were proposed. The projections based on the three scenarios were simulated for the year 2050. Five types of land use/land cover were identified and evaluated. They were primary forest, secondary forest, human settlements, areas without vegetation and water bodies. Results from the land use/land cover change analysis show a substantial gain for the secondary forest. The surface area of the primary forest was reduced from 55.8% in 1990 to 37.7% in 2017. Moreover, the three projected scenarios estimate further losses of the surface are for the primary forest, especially under the stationary and pessimistic scenarios. This highlights the importance and probably urgent implementation of conservation and protection measures to preserve these ecosystems and their services. Based on the accuracy obtained and on the models generated, results from these methodologies can serve as a decision tool to contribute to the sustainable management of the natural resources of a region.
Land use land cover (LULC) of city regions is strongly affected by urbanization and affects the thermal environment of urban centers by influencing the surface temperature of core city areas and their surroundings. These issues are addressed in the current study, which focuses on two provincial capitals in Pakistan, i.e., Lahore and Peshawar. Using Landsat data, LULC is determined with the aim to (a) examine the spatio-temporal changes in LULC over a period of 20 years from 1998 to 2018 using a CA-Markov model, (b) predict the future scenarios of LULC changes for the years 2023 and 2028, and (c) study the evolution of different LULC categories and investigate its impacts on land surface temperature (LST). The results for Peshawar city indicate the significant expansion in vegetation and built-up area replacing barren land. The vegetation cover and urban area of Peshawar have increased by 25.6%, and 16.3% respectively. In contrast, Lahore city urban land has expanded by 11.2% while vegetation cover decreased by (22.6%). These transitions between LULC classes also affect the LST in the study areas. Transformation of vegetation cover and water surface into built-up areas or barren land results in the increase in the LST. In contrast, the transformation of urban areas and barren land into vegetation cover or water results in the decrease in LST. The different LULC evolutions in Lahore and Peshawar clearly indicate their effects on the thermal environment, with an increasing LST trend in Lahore and a decrease in Peshawar. This study provides a baseline reference to urban planners and policymakers for informed decisions.
The objective of this research is to know the predictive capacity of three LUCC models in the Baja California peninsula, Mexico, generated in 2008. Through the cartographic update method, three maps of roofs and uses of the Soil since 2018. This was done with remote sensing tools, geographic information systems, use of statistical analysis (R) software and change detection (DINAMICA-EGO). Once the 2018 LUCC maps were obtained in 2018, the reliability of each map was evaluated. And, finally, the predictive models carried out were evaluated. The 2018 LUCC maps presented a reliability greater than 96% in the three locations. The predictions of the LUCC models made in 2008 were very close to those observed in 2018 in two of them, since in the town of Santo Domingo the assertiveness was 77% and in San Quintin 86%, while in Tijuana was only 35%. The methodology used is a proposal that helps to know the degree of certainty of the predictive models of LUCC and the generation of updated cartography.
The sensitivity of hydrological processes to the changed environment is of great concern. The integrated impacts of climate change and urbanization in the future have been assessed in a watershed in Northwest China through a multimodel approach based on the combined application of Generalized Watershed Loading Functions, the Long Ashton Research Station Weather Generator, and the Land Change Modeler. The results showed that both climate change and urbanization would lead to more watershed streamflow, and their combination would have synergistic effects on additional increases. In addition, there would be different seasonal distributions of streamflow with a greater proportion of runoff. These study results are helpful in supporting projects and/or decision-making processes for managers by providing more insights into the regional hydrological changes affected by climate change and urbanization. The proposed methodology of the combined multimodel approach may be applicable in other areas with similar conditions. HIGHLIGHTS
A multimodel linkage approach was proposed for future hydrological estimations.;
The individual and combined effects of climate change and urbanization were investigated.;
Both individual effects would increase streamflow and change its seasonal distribution.;
The combined effect had a synergistic mechanism for the additive increase of streamflow.;
There would be higher streamflow with more runoff and an earlier peak of snowmelt in future spring.;
The climate and Land Use/Land Cover (LULC) changes evince the considerable impact on water balance components by altering the hydrological processes. So, the present work focuses on the evaluation of the combined impact of both the climate and LULC changes along with and without water storage structures on water balance components of the Krishna river basin, India under present and future scenarios with the help of Soil Water and Assessment Tool (SWAT). Sequential Uncertainty Fitting algorithm (SUFI-2) was used for the model calibration and validation, which were carried out at the Vijayawada gauge station. The coefficient of determination (R²) and Nash-Sutcliffe efficiency (NSE) values obtained during the calibration period were 0.63 and 0.61, respectively, whereas, in validation, these values were found to be 0.61 and 0.56, indicates satisfactory results. The results showed that the model simulations and performance were significantly influenced by the presence of water storage structures, whereas the LULC changes were effective at the sub-watershed level. Future LULC maps of 2025, 2055, and 2085 were simulated from the Cellular Automata (CA) Markov Chain model, and they were used along with future climate projections to investigate its impact on water balance components. The climate model projects an increase of water balance components specifically, surface runoff, streamflow, and water yield, except for evapotranspiration in the future. Whereas, the future LULC changes may influence in offsetting the streamflow 20 to 30% reference to the observed flow. Thus, LULC changes were significantly influenced the model simulations; therefore, it is essential to consider the LULC changes along with climate scenarios in climate change studies. Overall, the surface runoff, water yield, and streamflow may increase by 50% under Representative Concentration Pathway (RCP) 4.5, and they may double under the RCP 8.5 scenario by the end of the century.
Abstract. Sintang Regency is one of the 12 regencies and two cities in West Kalimantan Province, Indonesia. The total area of this regency is approximately 2.2 million hectares (ha) with 59% of the area designated as state forest area which provides high biodiversity and environmental services for adjoining communities. Through multi-stakeholder scenario planning, the government of Sintang Regency committed to protect and preserve forest resources for long-term landscape planning and sustainable utilization. Scenario planning yielded two possible outcomes in 2030 called “green” and “business as usual” (BAU) scenarios. Under the green scenario, future development without deforestation and land permits complied with spatial planning while under the BAU scenario, future conditions will be the result of past conditions without interventions. This study aimed to analyze land-use change in the regency over the past ten years. By applying the Terrset Land Change Modeller (LCM) algorithm, this study predicted the land use and carbon stock change of both scenarios in 2030. Three steps to apply the LCM are by analyzing the changes based on past history, modeling the transition potential and predicting the changes. Time series data of land cover data from 2006 to 2016 were used for this analysis. The results indicated that a green scenario prevents to stop deforestation about 117,136 ha (more than 5%) compared to the BAU scenario. Furthermore, the green scenario prevents the emission of 5 million tons of carbon (tC) for the regency indicating that the multi-stakeholder scenario planning process can be an effective strategy to preserve land and forest resources and promote sustainable development planning. The green scenario requires to limit the expansion of plantation areas, which are only allowed inside the current cultivation license and permit areas.
The consumption of natural resources increased simultaneously with the growth of the world‘s
population, resulting in different changes in the environment. In the climate issue, particularly, the
climatic issue becomes more evident due to the increase in the number of extreme events such as
hurricanes, floods, prolonged droughts and atmospheric pollution. In relation to atmospheric pollution,
the history of the municipality of Cubatão, belonging to the Santos microregion, reveals its effects on
Atlantic Forest vegetation affecting its development and also the photosynthesis process that captures
carbon dioxide (CO2). Additionally the study area is historically characterized by events of mass
gravitational movements due to high precipitation rates during certain periods of the year. The two
factors, pollution and mass gravitational movements, interfere in the availability of ecosystem
services, such as the carbon stock that brings improvement in air quality and consequent well-being to
the population. In this context, the present thesis aims to propose a method of analysis that considers
the carbon stock as an additional criterion for the distribution of ecological ICMS resources. Inserted
in the discussion of ecosystem services this proposal intends to suggest the alignment of São Paulo
environmental policy with global environmental issues. The ecological ICMS up to the present
allocates a maximum of 1% of total ICMS collection to municipalities with preservation units (0.5%)
and / or water reservoirs (0.5%). It is proposed to increase the transfers to 1% only in the conservation
units criterion, but not only considering the existence of this territory but the capacity of carbon
storage that the natural vegetation has, so at the end a proposal of legislation was drafted. On the other
hand, the methodological development to arrive at the proposed legislation considered the elaboration
of basic mapping of the area of study, elaboration of mappings synthesis of socioenvironmental
vulnerability identifying areas of greater risk related to mass gravitational movements and the
ecohydrological mapping classifying the influence of the vegetation stability of the slope.
Subsequently, a projection of the distribution of ecological ICMS resources, land use and land cover
model and carbon stock for the year 2022 is carried out.
Keywords: Ecosystem services, Landslides, Susceptibility, Vegetation and climate.
In recent decades, rapid population growth and human improper activities accelerated deforestation. Reducing Emissions from deforestation and forest degradation (REDD+) has been introduced as a strategy for reducing deforestation in developing countries. Thus, identifying areas with high-deforestation is important for site selection of the REDD+ projects. Transition potential modeling is applied as a tool for deforestation simulation. Drastic land use changes in the Central Hyrcanian forests caused a substantial reduction in forests cover. In this research, forest cover changes of the Central Hyrcanian forests were examined. Then, transition potential modeling using of three empirical procedures included: multi-layer perceptron (MLP) neural network, logistic regression, and similarity weighted instance-based learning were performed. Model validation was examined using relative operating characteristic and figure of merit. Using multi-criteria evaluation, suitable areas for REDD+ projects were identified and site selection using zonal land suitability method was performed. The results showed that the Central Hyrcanian forest decreased about 188,607 ha during 1984–2014 and the MLP model obtained the best accuracy. This study provides a general framework for site selection of REDD+ projects and also showed that sites with different suitability were found for REDD+ projects.
ResearchGate has not been able to resolve any references for this publication.