Journal of Spatial Analysis

---

Trees in urban areas have survived in a wide variety of conditions and constrains, whether developing in natural or manmade habitats. Due to environmental constrains and stresses, urban trees rarely achieve their biological potentials. Indeed, some of trees, in small groups, could excel in terms of age, biomass structure and dimensions in urban areas. In definition, tree hazard includes entirely dead or dying trees, dead parts of harmed live trees, or extremely unstable or unsteady live trees, which could be in result of structural defects and disorders or other factors that have the high risk to threaten the safety of people or property in the event of a failure especially in urban green spaces. Although the pruning or other rehabilitation and mitigation program of trees is known as the one of the principal domains of green space management, it is still includes shortcomings in terms of models and methodologies to classify or prioritize hazardous trees which need to be treated timely. The main objectives of this study were to: (1) model old Sycamore failure hazard in urban green spaces to elucidate the general and defects tree factors affecting on failure hazard; (2) prioritize the impacts of model inputs (general and defects tree factors) on tree failure hazard using model sensitivity analysis and (3) determining the trend model output changes in respond to model variables changes.

The following types of data (target trees characteristics) were solicited for each target tree: (1) General features: Tree Height (TH), trunk Diameter at Breast Height (DBH), Butt Diameter (BD) at ground surface and Vertical Length of Crown (VLC) were calculated from measured girth. Crown Spread (CS) was measured as the average of two diameters of projected drip line of the tree canopy.

(2) Tree defects: Detailed evaluation of individual trees was made according to 6 key physical defects, namely Internal Decay (ID) in percent, Length of Cracks (LC) in m, Crown Defoliation (CD) in percent, and Degree of Leaning (DL).

(3) Sycamore failure hazard classification: Sycamore Failure Hazard Risk (SFHR) classification was the probability that an entire tree, or part of it, will break and fall within the first or second year after study. Considering results of tree regular monitoring after two years, the following classes of tree failure hazard were determined. 1. Extremely Hazardous: Tree failure in the first year. 2. Semi-Hazardous: Tree failure in the second year.

ANN has been recently developed for data mining, pattern recognition, quality control, and has gained wide popularity in modeling of many processes in environmental sciences and engineering. ANN learns by examples and it can combine a large number of variables. In this study, an ANN is considered as a computer program capable of learning from samples, without requiring a prior knowledge of the relationships between parameters. To objectively evaluate the performance of the network, two different statistical indicators were used. These indicators are Mean-Squared Error (MSE), Mean Absolute Error (MAE), and coefficient of determination (R²).

In this study, the year of Sycamore failure in urban ecosystems is evaluated using tree variables and artificial neural network to determine the most effective tree variables in SFHR in urban green space. Various MLFNs were designed and trained as one and two layers to find an optimal model prediction for the SFHR and variables. Training procedure of the networks was as follows: different hidden layer neurons and arrangements were adapted to select the best production results. Altogether, many configurations with different number of hidden layers (varied between one and two), different number of neurons for each of the hidden layers, and different inter-unit connection mechanisms were designed and tested.

In this research, 200 trees were totally selected, then general and defects tree variables were recorded in urban green space. Considering the aim of study, which is discovering the relation between general and defects tree variables with SFHR class for modeling, the year of tree failure, was recorded.  

In the structure of artificial neural network, general and defects tree variables were tagged as inputs of artificial neural network and SFHR class was tagged as output layer. Considering trained networks (the structure of optimum artificial neural network has been summarized in Table1), Multilayer Perceptron network with one hidden layer and 4 neurons in layer created the best function of topology optimization (Table2) with higher coefficient of determination which equals 0.87 for class 1 and 0.9 for class 2. Sensitivity analysis respectively prioritizes Crown Spread (CS), Vertical Length of Crown (VLC), Degree of Leaning (DL) and Butt Diameter (BD), which effect on SFHR in class1 (Fig1) and class 2 (Fig2).

The determined procedure of SFHR changes with CS changes in the region declares SFHR increase nonlinearly with an increase in CS. The determined procedure of SFHR changes with VLC changes o declares that SFHR increase nonlinearly with an increase in VLC of tree. The determined procedure of SFHR changes with DL changes in the region declares SFHR increase nonlinearly with an increase in DL. The determined procedure of SFHR changes with BD changes o declares that SFHR increase nonlinearly with an increase in BD of tree.

Nowadays, artificial neural network modeling in natural environments has been applied successfully in many researches such as water resources management, forest sciences and environment assessment. The results of research declared that designed neural network shows high capability in SFHR modeling which is applicable in green space management of studied area. Sensitivity analysis identified the most effective variables which are influencing SFHR. So, to identify hazardous trees in study area, we should pay attention to the CS of Sycamore trees as the variable with high priority in determination of SFHR. We believe that, in hazardous trees management in urban green spaces, we should pay attention to some modifiable factors of tree, which are CS and VLC, by timely tree pruning. We suggest urban green space manager to run SFHR model, for tree stability assessment, before decision making on hazardous trees.

---

Risks assessment of forest project implementation in spatial density changes of forest under canopy vegetation using artificial neural network modeling approach
 
Nowadays, environmental risk assessment has been defined as one of the effective in environmental planning and policy making. Considering the position and structure of vegetation on the forest floor, the main role of forest under canopy vegetation cover can be noted in attracting and preventing runoff in the forest floor and reducing subsequent environmental risks. The purpose of this article is forest under canopy vegetation density changes modeling considering forest ecosystem structure and forest management activities as an environmental risk. The main objectives of this study were to: (1) model forest under canopy vegetation density in forest ecosystem to elucidate the ecological and management factors affecting on under canopy vegetation density; (2) prioritize the impacts of model inputs (ecological and management factors) on under canopy vegetation density using model sensitivity analysis and (3) determining the trend model output changes in respond to model variables changes.
In this study, Land Management Units (LMUs) were formed in the region considering ecological characteristics of land. LMUs were mapped out based on Ian McHarg’s overlay technique by ARC GIS 9.3 software. Ecological factor classes of an LMU differ from ecological factor classes of adjacent LMUs (at least in one ecological factor class). The following types of data were solicited for each LMU:
(1) Ecological variables: Altitude or elevation (El), Slope (Sl), Aspect (As), soil depth (SD), Soil Drainage (SDr),Soil Erosion (SE), Precitipation (Pr), Temprature (Te), trees Diameter at Breast Height (DBH), Canopy Cover (CC), and forest Regeneration Cover (RC).
(2) Management variables: Cattle Density (CD), Animal husbandry Dsitance (AD), Road Dsitance (RD), Trail Dsitance (TD), logs Depot Dsitance (DD), Soil Compaction (SC), Torist impacts (To), Skidding impacts (Sk), Logging impacts (Lo), Harvested trees volume (Ha), artificial Regeneration (Re) and Seed Planting (SP).
(3) Forest under canopy vegetation density: The percentage of under canopy vegetation density in each LMU was estimated by systematic random sampling method. In each LMU, a one square meter sample was taken. The average percentage of under canopy vegetation density in sample units of each LMU was calculated and used in the modeling process.
ANN learns by examples and it can combine a large number of variables. In this study, an ANN is considered as a computer program capable of learning from samples, without requiring a prior knowledge of the relationships between parameters. To objectively evaluate the performance of the network, four different statistical indicators were used. These indicators are Mean-Squared Error (MSE), Root Mean-Squared Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R²).
Various MLFNs were designed and trained as one and two layers to find an optimal model prediction for the under canopy vegetation density and variables. Training procedure of the networks was as follows: different hidden layer neurons and arrangements were adapted to select the best production results. Altogether, many configurations with different number of hidden layers (varied between one and two), different number of neurons for each of the hidden layers, and different inter-unit connection mechanisms were designed and tested.
In this research, 129 LMUs were totally selected, then ecological and management variables were recorded in them. In the structure of artificial neural network, ecological and management variables were tagged as inputs of artificial neural network and the percentage of under canopy vegetation density was tagged as output layer. Considering trained networks (the structure of optimum artificial neural network has been summarized in Table1), Multilayer Perceptron network with one hidden layer and 4 neurons in each hidden layer created the best function of topology optimization with higher coefficient of determination of test data (which equals 0.857) and the lowest MSE and MAE (which are 0.866 and 0.736 respectively). Considering the results of sensitivity analysis, ecological and management variables like the forest canopy density, cattle density in forest, soil erosion and soil compaction respectively show the highest impact on forest under canopy vegetation density changes (Fig1).
 
Table1. The structure of optimum artificial neural network in forest under canopy vegetation density

Output Layer	First Hidden Layer	Network features
Linear	Hyperbolic tangent	Transmission Layer
Gradient descent	Gradient descent	Optimization Algorithm
0.7	0.7	Momentum
1	4	Number of Neurons
-0.9 up to 0.9	-0.9 up to 0.9	Normalization

 
Table2. The structure of optimum artificial neural network in test data

MSE	MAE	RMSE	R2	Data	The structure of network( the number of neurons)-epoch
0.716	0.678	0.846	0.931	Trainning	Tanh(4)-160
0.793	0.703	0.891	0.894	Validation
0.866	0.736	0.931	0.857	Test

 

 
Fig1. The results of sensitivity analysis of artificial neural network model
 
Nowadays, artificial neural network modeling in natural environments has been applied successfully in many researches such as water resources management, forest sciences and environment assessment. The results of research declared that designed neural network shows high capability in forest under canopy vegetation density modeling which is applicable in forest management of studied area. Sensitivity analysis identified the most effective variables which are influencing under canopy vegetation density.
So, to identify hazardous LMUs in study area, we should pay attention to the canopy density of LMUs as the variable with high priority in determination of under canopy vegetation density. We believe that, in hazardous LMUs in forests, we should pay attention to some modifiable factors of LMU, which is cattle density in forest, by timely plan for livestock elimination. The forest under canopy vegetation density assessment model, in forest projects impact assessment, could be a solution in decision making about forest plan structure and implementation of similar projects in similar locations. 
 
Keywords: Forest plan, Environmental impact assessment, Multilayer perceptron, under canopy vegetation, artificial neural network
 

---

Zoning the possibility of landslides downstream of Sanandaj Dam
1-Introduction
The purpose of this study is to select the best model and identify landslide risk areas in the downstream basins of Sanandaj Dam. Every year, mass movements in the region cause damage to roads, power lines, natural resources, farms and residential areas, and increase soil erosion. Kurdistan province, with its mostly mountainous topography, high tectonic activity, diverse geological and climatic conditions, has the most natural conditions for mass movements. According to the available statistics, this province is the third province in terms of landslides after Mazandaran and Golestan. (Naeri, &Karami, 2018). The Gheshlagh River Basin is a mountainous region with a north-south trend. In terms of construction land, it is located on the structural zone of Sanandaj-Sirjan. The study area with an area of 970.7 square kilometers is located downstream of Sanandaj Dam. The city of Sanandaj is being studied within the region. Due to the type of climate and morphological processes, effective parameters are provided for landslides in the geography of the region.
2-Methodology
In this study, 9 effective factors for landslides, including slope, slope direction, fault distance, road distance, waterway distance, lithology, land use and precipitation were used. Using Google Landsat 8 ETM satellite imagery, Google Earth software identified 237 slip points. Then, the coordinates of the slip points were transferred to the Arc GIS software and a map of the landslide distribution area in this environment was prepared. Also, in this study, 89 non-slip points were prepared for use in the training and testing stages of Persephone neural network inside slopes less than 5 degrees. Artificial neural networks are made up of a large number of interconnected processing elements called neurons that act to solve a coordinated problem and transmit information through synapses. Neural networks begin to learn using the pattern of data entered into them. Learning models, which is actually determining their internal parameters, is based on the law of error correction. In this method, by correcting the error regularly, the best weights that create the most correct output for the network are identified. The neurons are in the form of an input layer, an output layer, and an intermediate layer. ahp includes a weighting matrix based on pairwise comparisons between factors and determines the level of participation of each factor in the occurrence of landslides. In this model, a large number of factors can be involved and the weight of each factor can be obtained using expert opinion.
3-Results
According to the results of the high-risk class neural network model, which occupies 31% of the basin area, it is the widest risk zone in the region. The middle class also accounts for more than 29 percent of the area, followed by the low-risk class. The results of the AHP model show that the middle class, with 32% of the area, has the highest dispersion in the region, the low-risk class and then the high-class are in the next position. The AHP model was used to prioritize the parameters affecting the landslide. The parameters of slope, lithology and land use play the most important role in the occurrence of landslides, respectively, and have the least role for slope direction, distance from fault and height. The results of the models used are consistent with the reality of the region's wide-risk hazards, and high-risk areas based on the models used are mostly located in the west and southwest of the basin. These areas correspond to the mountain unit and the steep slope. Based on the results of AHP model, the impact of human factors in the occurrence of landslides is weaker than the natural factors of the region and human factors play a stimulating and aggravating role in primary factors. Five methods for error detection were used to evaluate the models used
4-Discussion and conclusion
 .Due to the sensitivity of unstable slopes in the region, any planning to change the use and construction that increases the weight of the load on unstable slopes should be done in terms of geomorphological and geological conditions of the region.
Keywords: hazard zoning, landslide, neural network, AHP. Sanandaj Gheshlagh Watershed
 

---

Risk modeling of plant species diversity and extinction in Sorkheh_hesar National Park
 
Zahra Mosaffaei¹, Ali Jahani^2*, ³MohammadAli ZareChahouki, ⁴Hamid GoshtasbMeygoni, ⁵Vahid Etemad
 
¹ Masters of Natural Resources Engineering, Environmental Sciences, College of Environment, Karaj
^*2Associate Professor, Department of Natural Environment and Biodiversity, College of Environment, Karaj.
³ Professor, Department of Restoration of arid and mountainous regions, University of Tehran, Karaj
⁴ Associate Professor, Department of Natural Environment and Biodiversity, College of Environment, Karaj
⁵ Associate Professor, Department of Forestry and Forest Economics, University of Tehran, Karaj
 
 
Abstract
Full identification of hazards and prioritizing them for non-harm to nature is one of the first steps in natural resource management. Therefore, introducing a comprehensive system of evaluation, understanding, and evaluation is essential for controlling hazards. This study aimed to model and predict environmental hazards following increased degradation in natural environments by ANN. Thus, 600 soil and vegetation samples were collected from inhomogeneous ecological units. Soil samples were prepared by strip transect method according to soil depth in four profiles (5, 10, 15, 20 cm). Vegetation samples were also collected using a minimum level method using 2 2 square plots according to the type, density, and distribution of vegetation. Sampling was done in two safe zones and other uses were modeled using ANN in MATLAB environment. The optimal model of multilayer perceptron with two hidden layers, sigmoid tangent function and 19 neurons per layer and coefficient of determination of 0.90. The results of sensitivity analysis showed that soil moisture content would be effective in decreasing biodiversity and flood risk as well as increasing the risk of extinction of endemic species in the region, and then the apparent and true gravity and soil porosity and distance from the road play a key role in the degradation of cover. Vegetation has increased flooding and extinction risk. Therefore, it is recommended that measures related to soil and vegetation restoration in this park be taken to reduce future damages as soon as possible.
 
Keywords: Modeling, Artificial Neural Network, Environmental Hazards, National Park, Vegetation
 

---

Simulation of Zanjan temperature trends based on climate scenarios and artificial neural network method

Abstract
Severe climate changes (and global warming) in recent years have led to changes in weather patterns and the emergence of climate anomalies in most parts of the world. The process of climate change, especially temperature changes, is one of the most important challenges in the field of earth sciences and environmental sciences. Any change in the temperature characteristics, as one of the important climatic elements of any region, causes a change in the climatic structure of that region. The summary of the investigated experimental models on climate change shows that if the concentration of greenhouse gases increases in the same way, the average temperature of the earth will increase dangerously in the near future. More than 70% of the world's CO2 emissions are attributed to cities. It is expected that with the continuation of the urbanization process, the amount of greenhouse gases will increase. According to the fifth report of the International Panel on Climate Change, the average global temperature has increased by 0.85 degrees Celsius during 1880-2012. Therefore, knowing the temperature changes and trends in environmental planning based on the climate knowledge of each point and region seems essential. For this reason, the present study simulates the daily temperature (minimum, maximum and average) of Zanjan until the year 2100.

Research Methods
The method of conducting the research is descriptive-analytical and the method of collecting data is library (documents). To check the temperature of Zanjan city, the minimum, maximum and average daily temperature data from Hamdeed station of Zanjan city during the period of 1961-2021 were used. The data of general atmospheric circulation model was used to simulate climate variables (minimum, average and maximum temperature) using artificial neural network and climate scenarios in future periods. The output variables in this study are minimum, maximum and average daily temperature. Therefore, three neural network models were selected. For model simulation, model inputs (independent variables) need to be selected from among 26 atmospheric variables. Therefore, two methods of progressive and step-by-step elimination were chosen to determine the inputs of the model. In these methods, climate variables that have the highest correlation with minimum, maximum and average daily temperature were selected. By using RCP2.6, RCP4.5 and RCP8.5 scenarios, variables were simulated until the year 2100. Markov chain model was used to check the possibility of occurrence of extreme temperatures of the simulated values.

results
According to the RCP2.6, RCP4.5 and RCP8.5 scenarios and the simulation made by the neural network model, it is possible that on average the minimum temperature will be 3.6 degrees Celsius, the average temperature will be 3.3 degrees Celsius and the maximum temperature will be 2.7 degrees Celsius. Celsius will rise. The monthly review of the simulated data for all scenarios and the observed data of the studied variables shows that the average minimum, average and maximum temperatures in January and February, which are the coldest months of the year, will increase the most and become warmer. While the average minimum temperature in August, the average temperature in April and the maximum temperature in October will have the least increase. According to the simulated seasonal temperature table based on all scenarios, it was found that the average minimum, average and maximum temperature observed with the maximum simulated conditions were 6.9, 5.5 and 5.4 respectively in the winter season, and 3.3 in the spring season. 4, 2.3 and 3, in the summer season it increases by 3.3, 3.4 and 1.4 and in the autumn season it increases by 4.6, 4.5 and zero degrees. The frequency of extreme temperatures observed in all three variables of minimum, average and maximum temperature for the 25th and 75th quartiles is less than the number of occurrences of extreme temperatures simulated in all three scenarios. Based on this, all three variables will increase and there will be fewer cold periods. An increase in night temperature and average temperature in winter season and maximum temperature in summer season will occur more than other seasons. The difference between day and night temperature will be less in autumn and summer. Also, all seasons, especially the summer season, will be hotter and the occurrence of extreme temperatures is increasing for the coming years.

Keywords: climate scenarios, simulation, extreme temperatures, artificial neural network, Zanjan



 

---