ارزیابی تغییرات فضایی- زمانی زیرساخت سبز شهری مبتنی بر الگوریتم درخت تصمیم گیری فرآیندهای فضایی مطالعه موردی: سیمای سرزمین تهران

گسترش سریع شهرنشینی به‌همراه تغییرات بی‌رویه در کاربری/پوشش اراضی سیستم سیمای‌سرزمین تهران باعث اختلال در الگوی ترکیب و توزیع زیرساخت‌ سبز شهری شده است. هدف از تحقیق حاضر تحلیل تغییرات فضایی-زمانی الگوی زیرساخت‌ سبز شهری تهران متاثر از فرآیندهای فضایی سیمای‌سرزمین در بازه زمانی 2030-1990 (4 دوره 10ساله) می‌باشد. رویکرد پژوهش حاضر، شناسایی: (1) تاثیر انواع فرآیندهای فضایی در تغییر الگوی ساخت سیمای‌سرزمین و (2) روابط فیمابین الگوی ساخت فضایی سیمای‌سرزمین و فرآیندهای بوم‌شناختی در شکل‌گیری ظرفیت‌ها و محدودیت‌های سرزمین شهری است. برای این مقصود از الگوریتم درخت تصمیم‌گیری در راستای شناسایی فرآیندهای فضایی و از سنجه‌های سیمای‌سرزمین در تحلیل تاثیر فرآیندهای فضایی بر تغییر الگوی ترکیب و توزیع سیمای‌سرزمین تهران بهره گرفته شد. داده‌های کاربری/پوشش اراضی مورد نیاز، از تصاویر ماهواره‌ای لندست (1990 تا 2020) به‌دست آمد. به‌منظور پیش‌بینی تغییرات کاربری/پوشش اراضی برای سال 2030 از مدل ca-markov استفاده شد. کمّی‌کردن سنجه‌های سیمای‌سرزمین در دو سطح کلاس و سیمای‌سرزمین در 4 دوره زمانی مورد نظر انجام گرفت. یافته‌ها نشان می‌دهند در سطح کلاس و در بازه زمانی بین سال‌های 1990 تا 2020، فرآیندهای فضایی حذف و قطعه قطعه‌شدن به‌ترتیب سبب کاهش تعداد و مساحت اراضی سبز و باز در الگوی ترکیب و هم‌چنین کاهش پیوستگی و پراکنش نامتعادل آن‌ها در الگوی توزیع زیرساخت سبز سیمای‌سرزمین تهران شده است. همین‌طور در تمامی دوره‌های زمانی، فرآیند فضایی تجمع در لکه‌های ساخت و ساز تکرار شده است. داده‌های پیش‌بینی برای سال 2030 نیز بیانگر تاثیر فرآیند فضایی حذف بر هر دو کاربری/پوشش اراضی سبز و باز زیرساخت سبز سیمای‌سرزمین تهران می‌باشد. در سطح سیمای‌سرزمین نیز در بازه زمانی مورد بررسی شاهد ساده‌تر‌شدن بستر سیمای‌سرزمین در نتیجه غلبه کاربری‌های ساخت و ساز هستیم. نتایج حاصل برای تعیین نقشه راه برنامه‌ریزی الگوی فضایی زیرساخت سبز شهری کاربرد دارند.

Evaluating spatialtemporal changes of urban green infrastructure using decision tree algorithm of spatial processes Case study: Tehran Landscape System

Extended AbstractIntroductionThe rapid expansion of urbanization along with irregular changes in land use and Landscape System of Tehran has disrupted the composition and distribution pattern of urban green infrastructure. The present study seeks to analyse the spatialtemporal changes of urban green infrastructure in Tehran Landscape System affected by the spatial processes of land use changes in the statistical period (4 decades of 1990 to 2030). To reach this aim, the present study has identified (1) the effect of spatial processes on the changing landscape pattern and (2) the relationship between the spatial pattern and ecological processes of landscape and its influence on the capacities and constraints of green urban infrastructure. Materials & MethodsThe present study has focused on the landscape system of Tehran and its 22 districts as the study area. The descriptiveanalytical study consists of following stages: 1) Classifying urban land uses in19902000, 20002010 and 20102020 statistical periods using Landsat satellite images: (in Envi 5.3, Google Earth and Arc GIS 10.2 software), 2) Modelling and forecasting land use changes in 2030 using integrated model of Markov chain, automated cells (CAMARKOV) and TerrSetsoftware), 3) Determining spatial processes of landscape changes via decision tree algorithm. 4) Quantifying landscape metrics of composition and configuration of landscape pattern (green, open & built patches) at both class and landscape levels in the mentioned periods (in Fragstate 4.2 software). Results & DiscussionMany environmental decisions presume that some types or composition of land use are preferred to others. It is assumed that the spatial arrangement of elements in a landspace mosaic controls its ecological processes. This proposition is known as the pattern/ process paradigm, and forms the central hypothesis of landscape ecology (a branch of science developed to study ecological processes in their spatial context). Ten spatial landscape processes are considered to reflect changes in various patterns of landscapes (aggregation, attrition, creation, deformation, dissection, enlargement, fragmentation, perforation, shift, and shrinkage). These processes actually change the spatial structure of urban landscape and affect the quality of ecological processes in Tehran Landscape System. To identify the spatial processes responsible for landscape pattern changes during a defined period of time, a decision tree algorithm was developed. Decision tree required the following input: area or size (a), the perimeter or edge length (p), and number of patches (n) in each landcover class. The decision tree algorithm applied on Tehran Landscape System has indicated that spatial processes of ’attrition’ and ’fragmentation’ have led to a decrease in the integration of green and open patches in this landscape system. Measuring LSI and IJI metrics in 19902030 statistical period at the class level has also proved the previously mentioned finding. Increased ENNMN and decreased PLAND of open and green patches during two periods of 19902000 and 20102020 due to the spatial process of ’attrition’ have also showed this decreased integrity over time. These conditions have reduced the resilience of Tehran atmosphere and its capability to absorb air pollution and also have resulted in the recent development of thermal islands in different urban areas. Moreover, the COHESION metric has reduced in green and open patches due to the spatial processes of ’attrition’ and ’fragmentation’ at the class level. At the landscape level, the value of SIDI metric has also decreased from 1990 to 2020 and the same trend will continue according to 2030 forecast. Spatial process of ’aggregation’ in constructed patches has resulted in a decrease of NP and PD at landscape level during 19902000 and 20102020. Findings indicate the effect of spatial process of aggregation on constructed lands (highrise buildings) in the northern (such as District 1) and western parts of the city (such as District 22) which has interrupted wind movement and air purification in Tehran. The values of LSI and ED has also decreased at the landscape level due to the ’attrition’ of open and green patches leading to a reduction in the heterogeneity order of urban landscape system. On the other hand, increased IJI value in 2020 and 2030 indicates increased turbulence in distribution and also increased fineness index of open and green patches in the landscape system of Tehran. ConclusionFindings indicate that spatial processes of ’attrition’ and ’fragmentation’ have resulted in a reduction in the number and area of green and open patches in the composition pattern and also decreased coherence at class level from 1990 to 2020. This has resulted in an unbalanced distribution of the patches in the configuration pattern of green urban infrastructure in Tehran. The spatial process of ’aggregation’ has been repeated during the statistical period in the constructed patches. Data forecasted for 2030 shows the impact of ’attrition’ on changes occurring in both green and open land use. The landscape is also getting more simplified due to the dominance of constructed land uses. Findings can be applied to determine a roadmap and plan the spatial pattern of urban green infrastructure.