نوع مقاله : علمی - پژوهشی
نویسندگان
1 استادیار پژوهشی، بخش تحقیقات منابع طبیعی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان گلستان، سازمان تحقیقات، آموزش و ترویج کشاورزی، گرگان، ایران
2 استادیار پژوهشی، بخش تحقیقات منابع طبیعی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان مازندران، سازمان تحقیقات، آموزش و ترویج کشاورزی، ساری، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Background and objectives: Plant Species Richness (PSR) plays an important role in forest ecosystem functions and services. Despite the fact that Fine Woody Debris (FWD) occupy a large volume of the temperate forest ecosystem in the north of Iran, they have received less attention than Coarse Woody Debries (CWD). FWD is a component of natural forests, in addition to increasing the productivity of forest trees, helping to trees regeneration, maintaining and increasing soil moisture and nutrients, and long-term carbon storage, contribute the enhanced function of newly developing microhabitats with an important function in plant understory richness. Therefore, not paying attention to FWD leads to wrong estimation of the total volume of woody debries and its key role on the performance of forest ecosystems. Therefore, it was considered for current study that Boosted Regression Tree (BRT) machine learning technique used to model the PSR in unmanaged forest stands. For this aim, Oak - hornbeam stand in Loveh forest in the east of Golestan province, Iran was selected for sampling.
Methodology: For this purpose, 30 sample plots 400 m2 (20 m × 20 m) were set in the study area, and the type and percentage of plant cover was recorded based on the Braun-Blanquet index. In the following, the number of plant species recorded in each sample plot was the basis for calculating species richness. In order to measure the volume of FWD, the alignment of the sides of the sample plot of 400 square meters was used as the basis of calculation. Therefore, in line with the sides of each sample plot and in the form of a linear transect with a total length of 80 meters (equal to the perimeter of each sample plot), FWD intersected with the transect were identified. The names of tree species of each of the FWD is specified, and according to the cross diameter with the transect, the FWD were placed in one of three diameter classes: 1 to 2.5 cm, 2.5 to 4.5 cm, and 4.5 to 7.5 cm. To measure the percentage of soil moisture and organic matter, the soil was taken from the center of each sample to a depth of 15 cm. Soil moisture percentage was obtained by using the difference between wet and dry weight of soil and also the amount of soil organic matter by Walkley-Black method in the laboratory. The gmb package in R programming language was used to fit the boosted regression tree model. This model is one of the methods that helps to improve the performance of a single model by using the combination of multiple models. Therefore, in this method, the combination of two algorithms "regression tree and classification" and "boosting" is used. It should be noted that in this research, in order to reach the optimal number of trees, the number of 1000 was used as the starting point. In this study, the amount of species richness in each plot as the response variable, and the variables of slope percentage, slope aspect, altitude, soil moisture percentage, soil organic matter percentage, average of total volume of fwd, average volume of fwd in decay class 1, average volume of fwd in decay class 2 and the type of fwd were considered as predictor variables.
Results: Based on the results, the initial model fitted in the number of trees 7700 showed the highest accuracy. However, due to the lack of influence of some variables in the model, the variables of slope, slope aspect, altitude, the type of fwd and the average volume of fwd in decay class 1, these variables were excluded from the model based on the deviation changes. And the model was refitted in the optimal number of trees of 7800. Based on the final model of the BRT, the highest amount of species richness was recorded with the increase of soil organic matter to > 2.15% and in a soil moisture percentage >30%. Furthermore, a high amount of FWD from the first diameter class and with the decay class 2 (rotten) led to an increase in plant richness in the studied area. In the present study, the adjusted R squared > 0.99 with the Root Mean Square Error (RMSE) < 0.039 shows the high accuracy of the BRT model.
Conclusion: Although fwd comprise a large part of woody debries in the forests of the north of Iran, no specific place has been considered for it in any of the statistical protocols. This matter has led to the fact that there is no specific estimate of the volume of fwd and its role on the forest ecosystem is neglected. Therefore, this study was conducted with the aim of modeling the effect of this important component on the richness of plant species in a broadleaf stand in Loveh forests of Golestan province. The findings of the research showed that by creating favorable habitat conditions, the fwd increase the abundance of plant species and maintaining this component is important in increasing the organic matter of the forest soil. The high rate of decomposition in fwd compared to cwd causes that while maintaining and increasing the soil moisture, the organic materials in the wood are available to the soil layer in a short period of time. In recent years, some experts in the field of natural resources have emphasized the collection of fwd and their use in cellulose industries. A large part of these opinions has also found a scientific basis in the shadow of the minimization of the volume and role of the fwd and the lack of sufficient information about this very important component. If the findings of this research clearly show that the collection and removal of fwd will have a negative impact on soil moisture and organic matter and thus on the richness of plant species. Conducting similar and additional studies by including soil nutritional variables as well as other plant indicators can significantly help in confirming or rejecting this result
کلیدواژهها [English]