Journal of Spatial Analysis

---

Investigating the threats of mangrove forests
with the help of remotely sensed data
 
Behzad Rayegani: Assistant Professor of College of Environment, Department of Environment
 
 
Mangroves are a group of trees and shrubs that live in the coastal intertidal zone. Mangrove forests are very important because they are known as natural heritage and crucial in protecting coastal ecosystems. Mangrove forests stabilize the coastline, reducing erosion from storm surges, currents, waves, and tides. The intricate root system of mangroves also makes these forests attractive to fish and other organisms seeking food and shelter from predators. So, they are ideal places to support the elements of seafood networks. However, these forests are in danger of degradation because of rapid population growth, poor planning and unsustainable economic development. In the process of regenerating an ecosystem, it is necessary to identify the precursors of the threat, to consider the means to eliminate these threats. Therefore, identifying the threatening factors of the mangrove forest ecosystem is the first step in the restoration and protection of the ecosystem.
This study aims to investigate the change and the destruction in Mangrove forests and to identify threatening forces in the Hara Protected Area. Remote sensing is now widely used in studies of ecosystem changes because its information is available for the past, and there are many highly-developed techniques for change detection through remote sensing. Therefore, in order to identify the threatening factors of mangrove forests, remote sensing techniques were used to identify changed areas during a 15-year period. Images of ETM+ and OLI sensor from 2001 to 2015 were collected in the Hara Protected Area (Khorekhoran International Wetland). Given that we have used the multiple-date remote sensor data in this study, it was necessary to use absolute atmospheric correction methods for radiometric harmonization of data. So, with the aid of the ERDAS IMAGINE 2014 software, the Atmospheric and Topographic Correction (ATCOR) model was applied to all data. Subsequently, due to the difference in radiometric resolution of the OLI sensor with the ETM+ sensor, the output of ATCOR of both sensors was stretched into 8-bit data in order to eliminate the existing divergence in radiometric resolution. Also, based on spatial information, one of the image of OLI sensor at the current time was corrected geometrically, and then other images were registered to this image to eliminate geometric errors. There are many ways to detect changes with the help of remote sensing data, but we used two widely used techniques in this study: 1) post-classification comparison; 2) Change detection techniques of Algebra. Totally four different change detection methods were applied to these images. Change detection techniques of Algebra image method include image difference, image ratio, regression and post-classification comparison were used. At first, with the knowledge of the studied area, by combining the two supervised and unsupervised classification (hybrid method), the pixels that were known as mangrove forests were identified in both time periods of study. Then pixels with decreasing trend were determined by post-classification comparison method. From the image of the mangrove forests with the logic of Boolean (OR), a mask of mangrove was obtained, which showed the areas of mangroves during the two periods. This mask was used to make the second group of methods for determining changes (Algebra method) applied to the data. By doing this, in all algebra methods, the histogram showed the normal distribution. Finally, the vegetation spectral indices were applied to the data and their coefficient of variation was obtained in the Boolean mask area. Among these indices, NDVI showed better performance, so the algebra operation was used for this index. Accordingly, areas with decrease, increase and no change trends were visited and then overall accuracy and kappa coefficients were determined.
The results showed that the method of post classification comparison has the highest accuracy in the monitoring of vegetation changes in mangrove forests. This method with a total accuracy of over 93% and a kappa of more than 0.9 showed the highest accuracy in the detection methods of the changes, therefore, in the final examination and prioritization of the regions, this method was used. The surveys showed that the smuggling of fuel due to pour gasoline into the water and camel grazing are the most important destructive factors in the mangrove forest. After determining the rate degradation in four regions, these regions were ranked in order to carry out reclamation and restoration projects.
In the case of intelligent use of the capabilities of remote sensing, one can easily identify the threatening factors of an ecosystem. In the case of mangroves, the only limiting factor is tidal conditions. It is therefore recommended that, as in this study, images are chosen to determine the changes that are in a same tidal state
 
 
Keywords: Remotely Sensed change detection, Image Algebra Change Detection, Post-classification comparison, Determination of thresholds
 
 
 
 

---

Investigation of the effects of Covid-19 pandemic on UHI in residential, industrial and green spaces of Tehran

 Abstract
Rapid urbanization in recent decades has been a major driver of ecosystems and environmental degradation, including changes in agricultural land use and forests. Urbanization is rapidly transforming ecosystems into buildings that increase heat storage capacity. Loss of vegetation and increase in built-up areas may ultimately affect climate variability and lead to the creation of urban heat islands. The occurrence of natural disasters such as flood, earthquake … is one of the most effecting factors on the changes in intensity of urban heat islands. So far, a lot of research has been done on how it is affected by various types of natural disasters such as floods, earthquakes, droughts and tsunamis.
Two major environmental challenges for many cities are preventing flooding after heavy rains and minimizing urban temperature rise due to the effects of heat islands. There is a close relationship between these two phenomena, because with increasing air temperature, the intensity of precipitation increases. Drought is also a phenomenon that is affected by rainfall, temperature, evapotranspiration, water and soil conditions. One of the major differences between drought and other natural disasters is that they occur over a longer period of time and gradually than others that occur suddenly. Another natural disaster is the tsunami, which increases the area of water by turning wetlands into lakes, thereby increasing the index of normal water differences, which has a strong negative relationship with surface temperature. Ecosystems in urban areas play a role in reducing the impact of urban heat islands. This is because plants and trees regulate the temperature of their foliage by evaporation and transpiration, which leads to a decrease in air temperature.
Applying the locked down of the Covid-19 pandemic since the spring of 2020 has led to the global restoration of climatic elements such as air quality and temperature. In this study, the effects of Covid-19 locked down on the intensity of urban heat islands due to the limitations in industrial activities such as factories and power plants and the application of new laws to reduce traffic in Tehran were investigated. In this regard, the Landsat-8 satellite taken from a part of Tehran city has been used.

Materials and Methods
In order to investigate the effects of locked down in the spring of 2020 on the intensity of urban heat islands; the status of UHI maps in Tehran during the same period of locked down in three years before and one year after has been studied. The proposed method in this paper consists of two main steps. The first step is to generate UHI maps using land surface temperature (LST), normalized difference vegetation index (NDVI) and land use / land cover map analysis. In the second step, in order to analyze the behavioral changes in the intensity of urban heat islands during locked down and compare it with previous and subsequent years, changes in the intensity of UHIs are monitored.
UHI maps consist of three classes of high, medium and low intensities urban heat islands, which are based on performing the rule based analysis on land surface temperature characteristics and normal vegetation difference index derived from Landsat-8 satellite images as well as land use / land cover map. LULC maps are produced by support vector machine classification method consisting of three classes of soil, building and vegetation. In order to calculate the spectral features used in the rule based analysis, atmospheric and radiometric corrections must first be made on the red, near-infrared, and thermal spectral bands of the image captured by the Landsat-8 satellite. Then, vegetation spectral indices including NDVI and PV indices are generated.

Disscussion of Results
The capability of the proposed algorithm in this paper is first evaluated in the whole area covered by satellite images taken from the city of Tehran, and then in three areas including residential, industrial and green spaces. The data used in this article are images taken by the OLI sensor of Landsat-8 satellite in the spring of 2017-2021.
In the first step of the proposed method, maps of urban heat islands are generated based on multi-temporal satellite images of Landsat-8 taken in the years 2017to 2021 in the MATLAB programming software. Then, by comparing pairs of UHI maps in each of the residential, industrial and green space study areas, the trend of changes in the intensity of UHI is analyzed and the effects of locked down application in 2020 are evaluated.
The results of changes detection in urban heat islands in the period under consideration in this study showed that the percentage of areas that are in the class of high UHI in 2020 due to locked down of pandemic Covid-19 compared to the average of three years before that is 55.71%, has a decrease of 17.61%. The percentage of areas in the class of medium UHI intensity in 2020 due to locked down compared to the average of three years ago, which is 39%, increased by 4.8%, and in 2021 this amount again has decreased to less than the average. Also, the percentage of low intensity UHI class in 1399 compared to the average of three years ago, which is 5.3%, has increased by 12.8%.

Conclusion
In this study, the effect of locked down application due to the Covid-19 virus pandemic, which was applied in Iran in the spring of 2020 is investigated on the intensity of  urban heat islands in a part of Tehran city and three selected areas with residential, industrial and green space. Detection of changes in the intensity of urban heat islands was done based on the post-classification method and on the UHI classification maps related to the years 2017 to 2021. In order to produce UHI maps, in addition to the land surface temperature, the amount of vegetation index and the type of land use / land cover class were also used in the form of a set of classification rules.
Comparing the results of the study areas of residential, industrial and green spaces, it is important to note that the rate of reduction of the area of UHI with high intensity in the residential area is 5.25% more than the industrial area and 6.1% more than the green space. However, the reduction of locked down restrictions in 2021 had the greatest effect on the return of the area of ​​the high UHI class and caused the area of ​​this class to increase by 23% compared to 2020. These results indicate the fact that restrictions on the activities of industrial units such as factories and power plants and the application of new laws to reduce traffic, despite the same weather conditions in an area have been able to significantly reduce the severity of urban heat islands.

 Keywords: Urban Heat Islands, Land Surface Temperature, Vegetation Index, Change Detection, Covid-19