Volume 16, Issue 3 (11-2022)
مرتع 2022, 16(3): 468-480 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Mohamadi M, Jafarian Z, Tamartash R. Prediction of Plant Species Boibiversity using Generalized Linear Model (GLM) and Boosted Regression Tree (BRT) in Eastern Rangelands of Mazandaran. مرتع 2022; 16 (3) :468-480
URL: http://rangelandsrm.ir/article-1-1053-en.html
URL: http://rangelandsrm.ir/article-1-1053-en.html
Department of Range Management, Faculty of Natural Resources, Sari Agricultural Sciences and Natural Resources University, Sari
Abstract: (1021 Views)
Background and objectives: Prediction of species richness and diversity patterns are used to develop conservation strategies for biodiversity under regional and global environmental changes. Since modeling the distribution of plant species can provide useful and important information about identifying and introducing potential habitats of plant species, and also few studies have been done in the field of modeling species richness using environmental variables in Iran, so the study of modeling Species richness is important and necessary in the management of vegetation, which is the aim of this research to help rangeland management.
Methodology: This study was conducted in the Sorkh Grieve rangeland with an area of 2620 ha. vegetation and soil sample were taken via random-systematic stratification method. On each slope a transect with a length of 100 meters of was established in the direction of the height gradient and 10 plots were placed along it. Totally 260 plots of 1 m2 were used. In each plot, the list of species, percentage of species cover, and percentage of litter, rocks and number of species were recorded. Three soil samples were taken from a depth of 0-30 cm along each transect and some physical and chemical properties were measured in the laboratory. Physiographic variables were determined in Arc GIS software. Climatic factors were collected for a period of 15 years. Predictive climate variables in this research included annual relative humidity, average annual temperature, and average annual rainfall. The physical and chemical characteristics of the soil such as the percentage of moisture, sand, clay, silt, pH, EC, organic carbon, nitrogen, phosphorus, and potassium were measured. Prediction of species diversity indices of plants was done with two methods: generalized linear model (GLM) and, enhanced regression tree model (BRT). Analysis of the importance of environmental variables for GLM and BRT models was done in the biomode2 package. R2 and RMSE coefficient of explanation were used to evaluate the models. The area under the curve (AUC) criterion was used to compare the performance of these models.
Results: The results of the GLM model showed that altitude, silt, average annual rainfall, nitrogen, and average annual humidity were the most effective environmental factors affecting species richness, respectively. The BRT model results show that the variables of height, soil acidity, average annual humidity, and clay are the most important in species richness, respectively. The results of the GLM model showed that among the richness and species diversity indices, the highest R2 related to the richness index was 0.33. Also, the most important variables affecting this index were nitrogen, acidity, electrical conductivity, rainfall, and humidity. The results of the BRT model showed that among the examined indices, the highest R2 was related to the Shannon diversity index of 0.37. Also, the most important variables affecting this nitrogen index were altitude, clay percentage. and humidity. In general and according to the results, it appears that the environmental parameters affecting richness by the GLM model include altitude, average annual rainfall, average annual humidity, nitrogen, and silt. For the BRT model were altitude, annual humidity, acidity and clay. The evaluation results of the two models showed that the area under the curve (AUC) of the GLM model was 0.61 and the BRT model was 0.72, which shows that the BRT model performed better in modeling species richness in the region.
Conclusion: The result of this research provides good information about the distribution of plant species and affecting environmental factors on their diversity and richness. It is suggested that rangeland managers use the results of this research as well as similar research and give spatial importance to the environmental factors affecting the distribution and richness of species.
Methodology: This study was conducted in the Sorkh Grieve rangeland with an area of 2620 ha. vegetation and soil sample were taken via random-systematic stratification method. On each slope a transect with a length of 100 meters of was established in the direction of the height gradient and 10 plots were placed along it. Totally 260 plots of 1 m2 were used. In each plot, the list of species, percentage of species cover, and percentage of litter, rocks and number of species were recorded. Three soil samples were taken from a depth of 0-30 cm along each transect and some physical and chemical properties were measured in the laboratory. Physiographic variables were determined in Arc GIS software. Climatic factors were collected for a period of 15 years. Predictive climate variables in this research included annual relative humidity, average annual temperature, and average annual rainfall. The physical and chemical characteristics of the soil such as the percentage of moisture, sand, clay, silt, pH, EC, organic carbon, nitrogen, phosphorus, and potassium were measured. Prediction of species diversity indices of plants was done with two methods: generalized linear model (GLM) and, enhanced regression tree model (BRT). Analysis of the importance of environmental variables for GLM and BRT models was done in the biomode2 package. R2 and RMSE coefficient of explanation were used to evaluate the models. The area under the curve (AUC) criterion was used to compare the performance of these models.
Results: The results of the GLM model showed that altitude, silt, average annual rainfall, nitrogen, and average annual humidity were the most effective environmental factors affecting species richness, respectively. The BRT model results show that the variables of height, soil acidity, average annual humidity, and clay are the most important in species richness, respectively. The results of the GLM model showed that among the richness and species diversity indices, the highest R2 related to the richness index was 0.33. Also, the most important variables affecting this index were nitrogen, acidity, electrical conductivity, rainfall, and humidity. The results of the BRT model showed that among the examined indices, the highest R2 was related to the Shannon diversity index of 0.37. Also, the most important variables affecting this nitrogen index were altitude, clay percentage. and humidity. In general and according to the results, it appears that the environmental parameters affecting richness by the GLM model include altitude, average annual rainfall, average annual humidity, nitrogen, and silt. For the BRT model were altitude, annual humidity, acidity and clay. The evaluation results of the two models showed that the area under the curve (AUC) of the GLM model was 0.61 and the BRT model was 0.72, which shows that the BRT model performed better in modeling species richness in the region.
Conclusion: The result of this research provides good information about the distribution of plant species and affecting environmental factors on their diversity and richness. It is suggested that rangeland managers use the results of this research as well as similar research and give spatial importance to the environmental factors affecting the distribution and richness of species.
Type of Study: Research |
Subject:
Special
Received: 2021/09/16 | Accepted: 2022/02/11 | Published: 2022/11/1
Received: 2021/09/16 | Accepted: 2022/02/11 | Published: 2022/11/1
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
