1- Department of Environment, Faculty of Natural Resources and Environment, University of Birjand, Birjand, Iran
Abstract: (453 Views)
Extended Abstract
Background: Detecting changes is one of the main factors in studying the relationship between human activities and the environment. Land use emphasizes the social aspect of land use. In other words, land use is the output of activities that humans perform according to their economic and social needs. On the other hand, modeling and predicting future changes are also necessary to understand the quantity and quality of future changes and developments. Detecting and predicting changes are essential to take care of an ecosystem, especially in areas with rapid and often unplanned changes in developing countries. Recently, remote sensing techniques and geographic information systems have been used in most applications of detecting changes in land use and cover and, as a result, monitoring land uses. Using remote sensing information and geographic information systems, future changes can be predicted from the ratio of land use and cover changes over time, and appropriate measures can be taken. In this study, two methods, Random Forest and Ca_Markov, were used to detect and predict land use changes in the Nehbandan watershed in the horizon of 2040. In this regard, the objectives of this study are to detect land use in the Nehbandan watershed using the Random Forest method and to simulate land use in the Nehbandan watershed using the Ca_Markov method over a 33-year period (2008-2040).
Methods: In this study, data from the TM sensor of Landsat 4 and 5 satellites from 2008 and Landsat 8 in 2013 and 2022 were used to determine and evaluate land use changes in the Nehbandan watershed. After preprocessing the images, the image classification operation was performed by applying the Random Forest algorithm in the Saga GIS software environment. The Random Forest method is a supervised classification method in which a set of trees is used for classification and decision-making. Then, using the probability matrix of land use class conversion, the land use map of the Nehbandan watershed was predicted for the year 2040 by applying the CA-Markov model in IDRISI Terrset software. Following previous studies, the VALIDATE tool in Idrisi software was used in this study to validate the CA-Markov model, which was validated by comparing the model with a real map.
Results: The results of monitoring changes in the Nehbandan watershed from 2008 to 2022 showed that vegetation and residential land use have changed by 4.5 and 6.51 square kilometers, respectively. Also, the CA-Markov model prediction results showed that the land use changes in 2040 in vegetation and residential uses will be 11.23 and 9.63 km2, respectively, compared to 2008. The prediction results of the area and percentage of uses in 2040 show that the largest changes will be in the area of vegetation. Considering the trend of land use changes in the Nehbandan watershed, an increase in vegetation related to agricultural use and, at the same time, a limited increase in residential areas is expected by 2040 in the future. Moreover, the residential use will increase in the form of filling within the urban use, and the growth of the city will be less on the margins and in its suburbs. This is mainly because Nehbandan County is small in terms of population and area in this watershed, the occupation of most people is agriculture, and the growth of the county has been accompanied by the growth of agricultural use, which has been the occupation of most people.
Conclusion: The results of this study clearly show that analyzing land use changes in the Nehbandan watershed using advanced methods, such as random forest and CA-Markov, is feasible and accurate. The relatively high accuracy of the random forest model for classifying satellite images confirms the results of other studies that have considered this method to be one of the most efficient methods for distinguishing diverse land covers. Accurate monitoring of these changes has been carried out from 2008 to 2022, with regard to vegetation and residential uses with high accuracy (vegetation and residential use have changed by 4.5 and 6.51 km2, respectively). The results obtained not only demonstrate the capabilities of these tools in location modeling and accurate analysis of land use changes, but also this capability helps environmental and urban officials to make the best possible decisions for appropriate land management in the future, and emphasizes the need for comprehensive policymaking and management to reduce the negative consequences of land use changes in the region. The results of the land use classification evaluation between the simulated and reference maps show that a standard coefficient of 90% has been achieved for the preparation of land use and land cover maps. Land use and land cover changes are considered a consistent and reliable basis for simulating future scenarios and confirm the ability of the CA-Markov model to simulate future land use changes. Therefore, this model can be used to simulate land use change in the study area up to the horizon of 2040.
Background: Detecting changes is one of the main factors in studying the relationship between human activities and the environment. Land use emphasizes the social aspect of land use. In other words, land use is the output of activities that humans perform according to their economic and social needs. On the other hand, modeling and predicting future changes are also necessary to understand the quantity and quality of future changes and developments. Detecting and predicting changes are essential to take care of an ecosystem, especially in areas with rapid and often unplanned changes in developing countries. Recently, remote sensing techniques and geographic information systems have been used in most applications of detecting changes in land use and cover and, as a result, monitoring land uses. Using remote sensing information and geographic information systems, future changes can be predicted from the ratio of land use and cover changes over time, and appropriate measures can be taken. In this study, two methods, Random Forest and Ca_Markov, were used to detect and predict land use changes in the Nehbandan watershed in the horizon of 2040. In this regard, the objectives of this study are to detect land use in the Nehbandan watershed using the Random Forest method and to simulate land use in the Nehbandan watershed using the Ca_Markov method over a 33-year period (2008-2040).
Methods: In this study, data from the TM sensor of Landsat 4 and 5 satellites from 2008 and Landsat 8 in 2013 and 2022 were used to determine and evaluate land use changes in the Nehbandan watershed. After preprocessing the images, the image classification operation was performed by applying the Random Forest algorithm in the Saga GIS software environment. The Random Forest method is a supervised classification method in which a set of trees is used for classification and decision-making. Then, using the probability matrix of land use class conversion, the land use map of the Nehbandan watershed was predicted for the year 2040 by applying the CA-Markov model in IDRISI Terrset software. Following previous studies, the VALIDATE tool in Idrisi software was used in this study to validate the CA-Markov model, which was validated by comparing the model with a real map.
Results: The results of monitoring changes in the Nehbandan watershed from 2008 to 2022 showed that vegetation and residential land use have changed by 4.5 and 6.51 square kilometers, respectively. Also, the CA-Markov model prediction results showed that the land use changes in 2040 in vegetation and residential uses will be 11.23 and 9.63 km2, respectively, compared to 2008. The prediction results of the area and percentage of uses in 2040 show that the largest changes will be in the area of vegetation. Considering the trend of land use changes in the Nehbandan watershed, an increase in vegetation related to agricultural use and, at the same time, a limited increase in residential areas is expected by 2040 in the future. Moreover, the residential use will increase in the form of filling within the urban use, and the growth of the city will be less on the margins and in its suburbs. This is mainly because Nehbandan County is small in terms of population and area in this watershed, the occupation of most people is agriculture, and the growth of the county has been accompanied by the growth of agricultural use, which has been the occupation of most people.
Conclusion: The results of this study clearly show that analyzing land use changes in the Nehbandan watershed using advanced methods, such as random forest and CA-Markov, is feasible and accurate. The relatively high accuracy of the random forest model for classifying satellite images confirms the results of other studies that have considered this method to be one of the most efficient methods for distinguishing diverse land covers. Accurate monitoring of these changes has been carried out from 2008 to 2022, with regard to vegetation and residential uses with high accuracy (vegetation and residential use have changed by 4.5 and 6.51 km2, respectively). The results obtained not only demonstrate the capabilities of these tools in location modeling and accurate analysis of land use changes, but also this capability helps environmental and urban officials to make the best possible decisions for appropriate land management in the future, and emphasizes the need for comprehensive policymaking and management to reduce the negative consequences of land use changes in the region. The results of the land use classification evaluation between the simulated and reference maps show that a standard coefficient of 90% has been achieved for the preparation of land use and land cover maps. Land use and land cover changes are considered a consistent and reliable basis for simulating future scenarios and confirm the ability of the CA-Markov model to simulate future land use changes. Therefore, this model can be used to simulate land use change in the study area up to the horizon of 2040.
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
