Rapid population growth, land-use change, and industrial and urban pollution are currently among the most serious human-caused problems in the natural environment. On the other hand, the change in land surface temperature is a factor in assessing the quality of habitat, which is modeled on land-use change. Therefore, it is necessary to examine the correlation between short periods of land-use change and surface temperature. This study shows the role of different land uses in the study area by examining the effect of land use on temperature parameters during the statistical period of 2013-2019 in the northern watershed of Ilam province with two mountainous-plain parts and industrial location despite oil and gas industries. Random forest and machining algorithm techniques were used to prepare temperature and land use maps from Landsat satellite data of ETM and MODIS sensors with a spatial accuracy of 30 and 250 meters, respectively, in Google Earth Engine. The results showed that during the six years, the level of water, rangeland, agricultural, and garden land uses have increased by 0.15, 3.87, and 3.42 percent, respectively, and other land uses such as forests and barren lands have decreased by 13.33 and 0.11 percent. The temperature algorithm using the LST index and creation of thermal islands showed that the average temperature of the region in the base period (2014) was equal to 31.02°C, which at the end of the study period reached 31.88°C. During the six years the temperature has risen by 0.68°C. According to the results, climate and land use are affected by each other and each shows its effects on the other.
Extended abstract
1. Introduction
The surface of the earth is constantly changing due to various human activities (Wu et al., 2021; Vijayana et al., 2021). These changes may, in the most severe case, be related to land use pattern as a region (Gupta and Chatterjee, 2021). Land-use change and land cover (LULCC) is a general term for human land-use change (Daatakulloa et al., 2021; Wang et al., 2020). However, nowadays, global warming has become a global issue. Therefore, during the last four decades, the rapid growth of urbanization has caused the change and evolution of natural phenomena. These changes eventually cause changes in the earth's surface (Zandi et al., 2019). The expansion of urbanization changes the energy received by the earth's surface, which can play a decisive role in meeting the needs of outdoor water and evapotranspiration in an area (Saher et al., 2021). Therefore, the destruction of green space and agricultural use has an important role in the formation of thermal islands (Zandi et al., 2019). Syahira et al. (2021) examined heat islands in an urban area due to land-use change and climate change in the Melbourne region of Australia and found that the thermal islands created were closely related to land change. Ren et al. (2021) showed that there are significant changes in urban and rural areas by examining satellite images of MODIS sensors for the relationship between thermal islands and urban land use in large Chinese cities. In other words, large urban areas of heat islands have expanded. The present study was conducted to investigate the effect of land use on temperature parameters in the northern watershed of Ilam province.
2. Materials and methods
The northern watershed of Ilam province is located in the west of Iran. This area is about 7195.81 square kilometers. In terms of geographical coordinates, the study area is located at 45° 40' to 46° 53' east longitude and 32° 53' to 34° 02' north latitude. In this study, to determine the most effective land use on temperature changes, thermal islands and land-use changes for a period of six years (2014 to 2019) were extracted. The study used Landsat ETM and MODIS satellite data from the US Geological Survey. To prepare the training points, classification algorithms were used and for accuracy testing, high-resolution images of Google Earth were used.
2.1. Random forest algorithm: Currently, one of the best learning algorithms is the random forest algorithm. A stochastic forest algorithm is a nonparametric machine learning algorithm based on a bunch of decision trees. A large number of decision trees grow in the RF algorithm classification (AboTalebi et al., 2017).
2.2. A backup vector machine algorithm: One of the capabilities of the backup machine is overcoming the problem of non-linear distribution of educational data. In this case, using kernel functions, data is transferred to a larger space in which better resolution is performed and the separating cloud screen is determined in that space (Ishaqi et al., 2016).
3. Results
In the study area during the period, water use, rangeland, agricultural and garden lands increased by 0.15, 3.87, and 3.42 percent, respectively, and for other uses such as forests and barren lands, respectively 13.33 and showed a decrease of 0.11%. Among the studied year, most changes are in barren and urban campuses and forest lands. According to figures, land-use changes annually, and unlike many studies that consider long periods for studying land-use change suitable, these figures indicate annual land use changes.
Temperature maps of the study area showed that the average temperature of the region in the base period (2014) was equal to 31.02°C, which at the end of the study period reached 31.88°C over six years, which is equal to 68. The temperature has increased by 0°C (Figure 3). In 2016, with the decrease of barren and urban lands, the temperature also decreased and reached 35°C. In 2014, the area of ​​agricultural land use was 660.15 square kilometers, the temperature was 30.96, and with the increase of these lands in 1398, the temperature decreased from 43.25°C to 31.25°C. The temperature has increased with the increase of rangeland lands, but on the contrary, this situation has occurred in forest areas and with the decrease of forest lands, the temperature has increased.
4. Discussion and Conclusion
According to the obtained maps, it was found that most changes are in barren and urban campuses and forest lands. Wu et al. (2021) and Gupta and Chatterjee (2021) stated that one of the most important changes is the increase in urban (residential) use. Khidmatzadeh et al. (2021) introduced the trend of urbanization and reduction of vegetation and garden use, but by examining the thermal islands created in the obtained temperature maps, it was found that the average temperature of the region in the base period (2014) is equal to 31.02°C, which at the end of the study period reached 31.88°C over six years, which increased by 0.68°C. Khidmatzadeh et al. (2021) also showed the relationship between the increasing residential area and temperature changes in Urmia. Shabani et al. (2019) also introduced land-use change as one of the reasons for the increase in temperature in Saqez. With the increase of barren and urban lands, temperature changes have increased, but this rate has not always been constant so the temperature in 1393 for barren land use with an area of ​​321.47 square kilometers was equal to 34.97°C, which in the following year decreased to 3112.67; the temperature has reached 35.26°C.