جستجوی مقالات مرتبط با کلیدواژه « رگرسیون خطی چندمتغیره » در نشریات گروه « جنگلداری »

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.

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.