جستجوی مقالات مرتبط با کلیدواژه « رگرسیون لجستیک » در نشریات گروه « اکولوژی »

Having accurate quantitative and qualitative information about the state of forest stands, is necessary for any basic management and planning, to reduce the effects of forest degradation. The current study aimed to model the destruction of Hyrcanian forests under the effects of density and volume (per hectare) variables, using logistic regression. In total, 252 plots of 1000 m2 area were measured. In each sample plot, species name, Diameter at Breast Height (DBH), height, density, volume and the presence or absence of forest degradation were measured and recorded. To model forest degradation, logistic regression model was utilized and Omnibus test, log-likelihood and pseudo r-square (Cox&Snell and Nagelkerke) coefficients were used to evaluate the  model. Results showed that the mean of density and volume of trees were 136.8 tree and 239.9 m3/ha, respectively. In addition, the results indicated that 46.82% of the study area was degraded. The results of correlation test showed that there was a srtong negative correlation between quantitative variables and the forest degradation. The independent variables of density and volume of trees were respectively explained 61.6 to 82.3% of the variance of the dependent variable (forest degradation). Among the input variables of the regression model, the effects of density and volume were significant on the forest degradation and it was possible to predict the changes of dependent variables (presence or absence of forest degradation).

Industrial and medicinal plants are one of the by-products of rangelands, the exploitation of which has a role in the economy of rangelands. The preservation and development of distribution of medicinal plants at the level of rangelands is important from various economic, social and technical aspects of rangelands. There are many industrial and medicinal plant species in the rangeland. Anchusa strigosa is a medicinal and indigenous species of Dezful. One of the important issues in the principled exploitation of these medicinal plants is to determine, its habitat areas. The purpose of this research is to investigate the relationship between the biophysical properties of the potential habitat of this species using different data mining techniques. Evaluating the effectiveness of data mining models, first 30 presence points and 30 absence points of the species were recorded using vegetation map and field studies. Then the parameters of height, slope, temperature, rainfall, texture and depth of soil were extracted from the obtained maps using geostatistical methods based on regional data. Determination of the influence coefficient of each of the mentioned parameters was done using the vector machine method. Then, the presence of the mentioned species was predicted based on biophysical parameters using data mining methods. The data mining algorithms used in this study include particle swarm, logistic regression, artificial neural networks and the nearest neighbor.The results showed that particle swarm algorithm and artificial neural network with accuracy rate of 0.95 and kappa coefficient of 0.9 have a higher accuracy in determining the species presence areas.The results of this study showed that using data mining methods, it is possible to predict the presence of the desired species, based on biophysical parameters with high accuracy.

Detection of climate parameters indicates that climate change has begun in Iran and it is necessary to identify the potential habitat of the indicator species, present and future years, under climate warning models. In this regard, it should be examined whether the increase in incident temperature in the country, will have a positive effect on the presence of species in the habitats of the distribution site, or a negative effect? For this purpose, the current and future range of Stipa barbata species for the next three decades (year-2050) was predicted under two climate warning models (scenario RCP4.5 and RCP8.5) using the logistic regression model. Output maps, with probability of occurrence between zero and one, to four categories unsuitable habitat (0-0.25), almost suitable habitat (0.25-0.5), habitat with high suitability (0.5-0.75) and habitat with very high suitability (0.75-1), were grouped. Finally, based on the coefficients of the variables in regression relationships, effective variables for the current and future range of S. barbata species were introduced. Based on the results; only temperature-related indices (BIO7 and BIO10) were found to be effective on the distribution of S. barbata species, present and future. The values of the indices increase with increasing climatic conditions. As a result, the range of S. barbata species decreases in response to climate change. This indicates during the next three decades, the presence of S. barbata species will decrease in the habitats of the region and there is a risk of its removal from the ecosystems of the region. In total, climate change and consequently increase in temperature characteristics causing the vertical spresd of S. barbata species, and it will move towards higher latitudes along the altitude gradient of the region.

The aim of this study was to evaluate the efficiency of logistic regression method in determining of the environmental factors effecting on the presence of Dorema ammoniacum D. DON, and to identify the habitat conditions of this species. Vegetation sampling was conducted random-systematically with establishment of plots along 9 transects in the presence region and 9 transects in the absence region. The size of plots was determined according to the type of existing species and using minimum area method and the number of plots (10 plots along each transect) was determined using statistical method. Data related to the presence and absence of the selected species (180 plot samples of each) was collected and environmental factors of the sampling sites including: topography, climate and soil were investigated. For soil sampling, profiles were drilled at the beginning, middle and end of each transect from two depths of 0-30 and 30-60 cm, then samples of each depth were mixed together and prepared as a composite sample, and the physical and chemical properties of the soil were measured at the laboratory according to the usual methods. Habitat prediction modeling was performed using logistic regression method and D. ammoniacum habitat predicted map was derived using the prediction model. According to the prediction model, slope of habitat, lime and clay of the second soil depth have the highest effect on the species distribution. After determining the threshold of optimal presence, the accuracy of the predicted map with the actual map was measured using the kappa coefficient. According to the optimal presence threshold (0.5), and the amount of kappa coefficient (0.5), the result of model evaluation was concluded as a good agreement. Results show that the logistic regression method is able to predict the habitat of D. ammoniacum, thus the distribution of this important and valuable species, and derived information from this study can be used in the locating of the areas prone to restoration of this species at the study area and the areas with the same ecological conditions.

There is closely interrelation between vegetation and environmental factors in a natural ecosystem. The aim of this study is applicability of analytical hierarchy process for this relationship. For this purpose was used logistic regression (that was done in it measurements of vegetation and environmental factors) as control method. First, study area was divided into 33 homogeneous units on slope, aspect and elevation. Then 26 designed questionnaires were filled for each site. Questionnaires were designed on basis three main criteria of soil, physiographic and climate and sub-criteria included soil, physiographic features and climatic characteristics. Then according to the presence of studied plant species in the criteria and sub-criteria classes, scoring was done for the environmental factor. Effective factors were prioritize and prepared for use in Expert Choice for each plant species separately and also incompatibility coefficient of each questionnaire was calculated.10 sampling plot were established and two soil samples were gathered from two depth of 0-15, 15-30 cm in each sampling site. 14 soil characteristics, 3 physiographic factors and 5 climate factors were measured. Most important factors affected on dominant species were determined with these environmental data with using logistic regression and were assessed using total accuracy of classification and correlation coefficient. The results of Logistic regression showed that the environmental factors studied, the soil has a significant role in the distribution of species and physiographic and climatic factors in subsequent courses in the degree of importance. Also AHP method is able to well represent this connection. If experts have sufficient knowledge about of region, it is suggested to use the method of less costly and time-consuming AHP than else methods that require more cost and time.

Habitat suitability prediction of important plant species in formal and indigenous Knowledge is known as one of the considerable factors in rangelands reform and development, since one of the main points of success or failure of these activities is adaptation with needs and culture of the area residents. This survey is about feasibility study of preparing prediction map of the possibility of presence of some range species being important for natives, local experts and having Multifunctional value. This study was conducted in Dilegan rangeland in Kohgiluye and Boirahmad province. Important species were identified based on interviews with local experts, scientific centers associated with natural resources and tribal affairs. Then logistic regression model was used for habitat prediction map of the selected species and regarding the purpose of study, vegetation data and environmental factors were collected in the rangeland. By creating relationships between environmental variables and the presence or absence of species and applying these relationships in GIS, habitat prediction map of Astragalus gossypinus Fischer, Prangos ferulacea (Schlecht) Boiss. and Ferulago angulata (L.) lindi was created in the ArcGis version 3.9. Kappa (κ) was used to evaluate the prediction habitat maps. Results showed that gradiant analaysis of environmental variables by logistic regression can be used as one of the appropriate methods for preparing habitat suitability prediction map of plants. In this method, by determining the presence or absence of plant species, its relationship with environmental factors is identified, and in other areas, in case of knowing environmental factors, predict the possibility of presence of plant species.

Traditional soil maps generally consist of some polygon units delineated on the basis of soil variation uniformity. Regarding the difficulties of this kind of maps، researchers have recently tried to use digital soil mapping (DSM) techniques in order to present spatial distribution of soil properties with less cost and time. In this research، five DSM techniques were applied to predict soil classes in the Yazd-Ardakan plain. Firstly، the USDA soil classification of 422 soil profiles across the Yazd Province were determined and coded. Then، geomorphometric parameters were related to soil data. Results showed that elevation، MrVBF، and wetness index have the most influence on prediction of soil classes up to great group level. Moreover، results indicated that decision tree، genetic algorithm، logistic regression، artificial neural network and discriminant analysis predict soil classes with overall accuracy of 60. 5%، 49%، 40%، 37% and 31%، respectively. Results also confirmed the capability of decision tree. Later technique could increase the accuracy of prediction up to 48% compared to the discriminant analysis technique. Therefore، using of decision tree model with geomorphometric parameters for soil mapping and updating soil maps is suggested in future studies.

Restoring high value medicinal and industrial plant species requires identifying habitat potentials of the target species. This study aimed to generate Astragalus verus potential distribution map in Fereydunshahr rangeland، Isfahan Province. First، soil maps were obtained from 70 soil profiles and climatic maps were produced from 10 climatic stations data. Geostatistical methods were used to convert the obtained data to digital maps with 90 by 90 meter resolutions. Finally potential distribution map of the species in the study area، was produced using presence-absence data of 100 sites and performing Logistic regression method. According to results، annual mean precipitation، clay، mean temperature of warmest season and slope were the most important environmental factors influencing the distribution of the species. According to model output، 21 and 27 percent of the whole study area had excellent and fair habitat condition for the species، respectively. The produced species distribution model had good accuracy as calculated Kappa coefficient and ROC area under curve plots were 0. 75 and 0. 5، respectively. This modeling approach enables to allocate appropriate areas to protect and restore valuable native plant species such as Astragalus verus.