multiple linear regression

Wildfires represent a significant environmental challenge globally, with climate change exacerbating their frequency and severity in many regions. In the Iranian Northern Zagros Forests, a Mediterranean climate prevails, characterized by hot, dry summers and relatively mild winters, rendering these forests particularly susceptible to fire. Understanding the relationship between climate change and wildfires in this region is essential for informing effective management strategies and mitigating associated risks. This study aims to investigate the effects of climate change on wildfires in the Iranian Northern Zagros Forests, focusing on changes in fire frequency and extent over a 17-year period from 2006 to 2023.

To achieve this objective, we employed a combination of statistical analyses, including Pearson's correlation and multiple linear regression (MLR), using climate data and historical wildfire records. We analyzed the relationship between the number and extent of fires and climatic variables such as average annual temperature and relative humidity. Pearson's correlation was utilized to assess the strength and direction of the linear relationship between these variables, while MLR was employed to predict the burned area based on temperature and humidity.

From 2006 to 2023, the forests of Sardasht experienced a significant number of fires, with the most occurring in the months of August, July, September, and June. The greatest scorched area of 211.62 hectares was reported in 1402, while the most frequent fires were recorded in 1400. This tendency is likely to continue. The study found a positive and significant relationship at the 95 percent confidence level between the number of fires and the burned area in Sardasht forests and the average annual temperature, but no significant relationship was found between other climate data and fire factors. The results of the MLR analysis demonstrated the predictive power of temperature and humidity in determining the burned area, highlighting their significance as drivers of wildfire activity in the region.

In conclusion, our findings underscore the escalating threat of wildfires in the Iranian Northern Zagros Forests driven by climate change-induced changes in temperature and humidity. The observed increase in fire frequency and extent highlights the urgent need for proactive measures to mitigate fire risks and enhance resilience in fire-prone ecosystems. By integrating scientific research with stakeholder engagement and policy action, we can develop effective wildfire management strategies that balance ecological conservation with socio-economic priorities, thereby ensuring the long-term viability of forest ecosystems and the communities that depend on them.

As a sustainable forest indicator, biodiversity plays a crucial role in understanding the effects of climate change on forest ecosystems. Measuring the diversity of trees and shrubs in forests is essential for monitoring and evaluating changes in biodiversity. Remote sensing (RS) is an effective tool for collecting such data. To estimate tree and shrub species diversity, we used Sentinel-2 data from August 10 and October 13, 2021. We measured 75 field plots with dimensions of 20 m × 20 m in the Patom, Namkhaneh, and Gorazban districts. In each field plot, the tree species and diameter at breast height of all trees with a diameter greater than 7.5 cm were recorded. We used the Jaccard and Sorensen indices in R software to calculate the beta diversity indices for each sample plot. Preprocessing steps were applied to the Sentinel2 data, and we then performed several spectral transformation approaches, that is, vegetation indices (VIs), principal component analysis (PCA), and Tasseled Cap, and generated texture variables. A vector map was used to extract the spectral and textural values corresponding to each field plot. Correlation analysis between the measured species diversity and spectral and textural variables was conducted at a 95% probability level. Multiple Linear Regression (MLR) analysis was performed using stepwise and Random Forest (RF) methods for modeling. Our regression analysis revealed that texture variables with a window size of 5×5 and spatial resolution of 10 m in Sentinel-2 summer images had the best performance in estimating the Sorensen diversity index( R2= 0.383 and RMSE%= 36.57). However, based on our results, we can conclude that the Sentinel-2 data has a moderate performance in estimating diversity in the Patom, Namkhaneh, and Gorazbon districts.

The humus index is a numerical score based on the visual assessment of organic horizons and humus forms identification. This index reflects different characteristics of forests stand and soil. Forest humus is an intermediate between plants and soil, which is affected by various factors, such as topography, climate, vegetation and soil. Nevertheless, by digging the humus profile and determining its index, it is also possible to determine many ecological factors. The aim of this study was to investigate the humus index and its relationship with topographic, soil and forest cover characteristics along an altitudinal gradient in the four-altitude class, including 0, 500, 1000 and 1500 meters above sea level (m a.s.l.), in Vaz forests located in Mazandaran province.

In each altitude class, the characteristics of the trees such as crown canopy, density, diameter and height, as well as topography and soil characteristics such as altitude, slope percentage, soil moisture and temperature were taken in 400 m2 plots, and in each plot, five 30 × 30 cm profiles for identification of humus forms were drilled and then identified and classified based on Humusica classification. In order to evaluate the humus index between different altitude classes was used one-way analysis of variance (ANOVA), as well as to analyze its relationship with the mentioned characteristics, multiple linear regression analysis (MLR) was used.

The results showed that the highest amount of humus index and litter thickness belonged to the altitude class 1500 m a.s.l. and their trend is increasing along the altitudinal gradient. Also, the results of MLR indicate the existence of a linear and significant relationship is between humus index and topographic and soil characteristics with high determination coefficient (R2 = 0.76), especially soil temperature with (P-value = 0.000), while the tree cover characteristics with a low determination coefficient (R2 = 0.31), cannot well be a good predictor for evaluating the humus index. Meanwhile, soil temperature and trees crown canopy are two factors that have shown the highest t-statistics in relation to the humus index, so that the decrease in soil temperature and trees crown canopy has increased the humus index.

In general, it can be concluded that with increasing altitude, the organic part of the forest floor and Humus index increase, and their variability is mostly due to changes in topographic and soil temperature characteristics.

Canopy cover is an important forest structure parameter with many applications in ecology, hydrology and forest management. This study aimed at investigating the capability of Landsat 8 satellite OLI data for modeling and estimation of forest canopy in a part of the Zagros forests. First the images were evaluated in terms of geometrical and radiometric errors. Required processing such as Vegetation Indices, Principle Component Analysis (PCA), and Tasseled Cap transformation were performed on the images of the study area. To measure aboveground canopy cover using hemispherical photography method, 60 sample plots were designed with a square dimension of 45 × 45 m. The spectral values of the corresponding sample plots were extracted using a polygon map of sample plots. Forest canopy cover was modeled using stepwise multiple linear regression and the accuracy of the model was evaluated via the k-fold cross validation technique. The results of multiple linear regression between canopy cover with main and computational bands showed that the model obtained from SR vegetation index and band 8 with R2 = 0.662 and RMSE (%) = 15.24 was the best model. Overall, this study demonstrated that the estimation of forest canopy cover is cost-effective and requires low computation power using Landsat 8 satellite.

In the present study, firstly, the effective factors of Golestan national park forests fire have been identified. Then, by using these factors, forest fire development modeling is performed based on cellular automata. Multiple linear regression and genetic algorithm are used to determine the effective factors on Golestan national park forests fire. In order to investigate the effect of spatial resolution of the maps are used on the results of modeling, effective factors have been generated in different spatial resolution and these data are used as the input of the proposed algorithm. Also, the neighboring filters 3×3, 5×5, and 7×7 are used to investigate the effect of the neighboring filter in the forest fire development process. Cellular automata is used for modeling Golestan national park forests fire development, and the artificial bee colony is proposed to calibrate it. The results of this study show that using the proposed algorithm with 3×3 neighboring filter is more accurate than the other neighboring filters. In the best case, the Kappa index, the overall accuracy, and the relative operating characteristic are 0.924, 0.960, and 0.494, respectively that these results are for spatial resolution of 30 meters on November 17, 2010.

Knowledge of seasonal fluctuations is of great importance for sale management. The aim of this research is to assess the impact of sale season on the log and lumber prices. Hence، after evaluation of sale documents in Azarood forestry plan، the data of 410 timber sale lots at roadside of four seasons from 1992 to 2008 were extracted and incorporated into a dataset. Calculating the average selling price for each species group and products per season and year، four time series of seasonal price for two species group and products (log and lumber) were obtained. Then، all prices were deflated to the base year of 2007. The Augmented Dicky-Fuller test were then employed to evaluate stationarity of time series. Impact of season sale variables (independent dummy variable) on timber price was analyzed using multiple regression analysis، MRA، and stepwise method of SPSS 14. 0 software. The impact of sale season on log and lumber prices was appeared to be significant for first species group while insignificant impact was found for second species group (at 5% level). The results revealed that selling lumbers and logs of first species group in summer and autumn caused a price raise of 13 and 11 percent (i. e.، 120409 and 150798 Rial/m3 at constant prices of 2007) compared to other seasons، respectively. Wood trade and building boom in summer and autumn seems to contribute to timber prices raise in these seasons. Implications of such analysis seem to be of interest both for forest managers and wood industries.