Volume 18, Issue 1 (7-2024)
مرتع 2024, 18(1): 42-56 |
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:
Molaei M, Ghorbani A, Moameri M, Motamedi J, Hazbavi Z. Modeling Artemisia austriaca Habitat in Ardabil Province Rangelands. مرتع 2024; 18 (1) :42-56
URL: http://rangelandsrm.ir/article-1-1218-en.html
URL: http://rangelandsrm.ir/article-1-1218-en.html
Department of Range and Watershed Management, Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil
Abstract: (729 Views)
Background and objectives: Due to the need to determine the suitability of the habitat in providing emendation suggestions in biological programs, predicting the optimal habitat of important and valuable plant species in local and academic knowledge is considered one of the important matters in the rangeland's improvement and development. Nowadays, different modeling methods have been considered in this field. In addition to finding important influencing factors in the establishment and distribution of the species, this can determine its preferential tendency towards environmental factors. Therefore, this study was carried out with the aim of preparing a habitat prediction map of the Artemisia austriaca species, which is important from the point of view of both local people and experts of the region and has multi-purpose values, at the level of rangelands of Ardabil province with machine learning methods.
Methodology: In the Rangelands of Ardabil province, 675 sampling sites from the presence and absence of studied species were considered from 2018 to 2021. Two categories of environmental factors including bioclimatic variables and topographic variables, including primary and secondary topographic indicators were investigated in relation to the presence of the species. Maps of all environmental factors were prepared with 70% of the data and overlapped in geographic information software. Predicting the presence of A. austriaca with four methods; Generalized Linear Model (GLM), Generalized Cumulative Model (GAM), Random Forest Model (RF), and generalized boosted Regression Model (GBM) were performed in R software environment. The analysis of the importance of environmental variables for the models was done in the Biomode2 package. To evaluate the models, 30% of species data and three statistics of the area under the curve (AUC), kappa, and true skill statistic (TSS) were used.
Results: The modeling results showed that the precipitation variable in the coldest season (Bio19) was the most effective variable in the spread of A. austriaca species in all four studied methods. In the GLM method, the variables of seasonal temperature (Bio4) and topographic position index (TPI) were also found to be effective factors in the presence of the A. austriaca species. In the RF model, respectively, Bio19 and seasonal precipitation (Bio15), and in the GBM model, the variables Bio19, precipitation of the wettest month (Bio13), and Bio15 were introduced as the most important variables in the presence of the species. Also, in the GAM model, the results showed that the variables of altitude above sea level, Bio19, Bio15, and Bio4 are the most important in the distribution of the species in the A. austriaca. Comparing the performance of the models showed that the GBM model with AUC 0.97, Kappa index 0.8, and TSS 0.886 is the best model among the studied models, followed by the random forest model with the index under the curve 0.96, Kappa 0.79, and TSS equal to 0.86 is the second model approved in this connection.
Conclusion: The results of this research showed that the studied models, especially the GBM and RF models, are able to predict the optimal habitat of the A. austriaca in the rangelands of Ardabil province with acceptable accuracy. Based on the results of the mentioned models, climatic factors have a greater effect on the occurrence of A. austriaca in Ardabil province. The use of the results of this and similar research is essential in preparing the habitat identification of any plant species and suggesting suitable native species for the improvement of rangelands. Finally, using the environmental factors of each region, the probability for the success or failure of the establishment of plant species can be predicted. Because one of the main conditions in the success or failure of such operations is their adaptation to the needs of the suggested species in that area.
Methodology: In the Rangelands of Ardabil province, 675 sampling sites from the presence and absence of studied species were considered from 2018 to 2021. Two categories of environmental factors including bioclimatic variables and topographic variables, including primary and secondary topographic indicators were investigated in relation to the presence of the species. Maps of all environmental factors were prepared with 70% of the data and overlapped in geographic information software. Predicting the presence of A. austriaca with four methods; Generalized Linear Model (GLM), Generalized Cumulative Model (GAM), Random Forest Model (RF), and generalized boosted Regression Model (GBM) were performed in R software environment. The analysis of the importance of environmental variables for the models was done in the Biomode2 package. To evaluate the models, 30% of species data and three statistics of the area under the curve (AUC), kappa, and true skill statistic (TSS) were used.
Results: The modeling results showed that the precipitation variable in the coldest season (Bio19) was the most effective variable in the spread of A. austriaca species in all four studied methods. In the GLM method, the variables of seasonal temperature (Bio4) and topographic position index (TPI) were also found to be effective factors in the presence of the A. austriaca species. In the RF model, respectively, Bio19 and seasonal precipitation (Bio15), and in the GBM model, the variables Bio19, precipitation of the wettest month (Bio13), and Bio15 were introduced as the most important variables in the presence of the species. Also, in the GAM model, the results showed that the variables of altitude above sea level, Bio19, Bio15, and Bio4 are the most important in the distribution of the species in the A. austriaca. Comparing the performance of the models showed that the GBM model with AUC 0.97, Kappa index 0.8, and TSS 0.886 is the best model among the studied models, followed by the random forest model with the index under the curve 0.96, Kappa 0.79, and TSS equal to 0.86 is the second model approved in this connection.
Conclusion: The results of this research showed that the studied models, especially the GBM and RF models, are able to predict the optimal habitat of the A. austriaca in the rangelands of Ardabil province with acceptable accuracy. Based on the results of the mentioned models, climatic factors have a greater effect on the occurrence of A. austriaca in Ardabil province. The use of the results of this and similar research is essential in preparing the habitat identification of any plant species and suggesting suitable native species for the improvement of rangelands. Finally, using the environmental factors of each region, the probability for the success or failure of the establishment of plant species can be predicted. Because one of the main conditions in the success or failure of such operations is their adaptation to the needs of the suggested species in that area.
Type of Study: Research |
Subject:
Special
Received: 2023/08/24 | Accepted: 2023/10/14 | Published: 2024/07/31
Received: 2023/08/24 | Accepted: 2023/10/14 | Published: 2024/07/31
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
