عنوان مقاله [English]
Proper distribution of urban green space is one of the most important issues in urban planning and especially in management of urban green space. In other words, the physical expansion of cities destroys surrounding natural environments and arable lands. It also results in fundamental changes in ecological structure and functionof urban landscape, along with gradual changesin spatial structure and patterns of this landscape (Wang et al., 2008). Since ecosystem processes depends on its structure, landscape metrics have been accepted as a very useful tool for expressing the structure of urban green space and its human-causedchanges (Hessburg et al., 2013).There has always been discussion onacceptable per capita green space or changes in green space over time and place. Iranian cities are no exception in this regard, thougheven a city enjoying a high ratio of green space per capita may still lack enough green space per capita in some districts. This suggests the necessity of investigating various measures and avoiding studies limited to per capita green space and urban forestry. (Botequilha and Ahren, 2002). If as an ecological structure,green space is proportional to populationcomposition and distribution, ecological performance and land use type of an urban area, it can have important ecological functions. Since most studies on urban green space have primarily focused onfinding a proper location, calculating appropriate per capita green space and introducing suitable species for green space, investigatingthe spatial distribution of urban green spaceseems to be of great importance. Therefore, the present study seeks to investigate the spatial pattern and distribution of public green space in Khomein using a landscape approach.
Materials and methods
The study area, Khomein, is bounded by agricultural lands and gardens in its northeast, west, and partly in its south. Only the main area of urban texture is located on barren lands (Abbasi et al., 1986). The study area includes four districts of Khomeinin which the pattern of green space distribution isinvestigated.
Sentinel-2 images were used in the present study. Satellite images were processed and then, their geographical effects were extracted inthe first step of classification. Different indices were defined for each patch of the image and using supervised method, images were classified into four classes of agricultural lands, barren lands, urban parks and residential areas in accordance with the training data. Visual method was used to improve classification results. In this method, classification results are matched with the imagesand possible errors are rectified. Google Earth was used to evaluate the accuracy of results obtained from classification of satellite images. In the next step,the base map of the present study was produced and then, the layer containing urban parkswas integrated with the layers prepared for four districts of Khomein. It should be noted that the present study focuses on urban parks prepared by the municipality for public use and does not include other urban green spaceareas such as the green belt or private gardens, etc.
To study the spatial distribution of green space, measures of land cover were calculated and analyzed in each of the four districts. Geographic Information System (GIS) and Landscape Measurement Analysis Program (FRAGSTATS) were among the tools used to calculate and measure landusein the present study. Landscape metrics used in the present study included:
Landscape Shape Index (LSI) which measures the area of the largest patch in a class divided by the total area of that landscape (multiplied by 100 to convert to percent)
Euclidean Nearest Neighbor distance (ENN) which is the average distance between patches in a class. Meter is used as the standard unit of measurement for this index.
Perimeter /Area Ratio (PARA) which is the ratio of the perimeter of the patch (m) to its area (m2). This measure lacks a specific unit and for PARA> 0 it is without a limit.
Number of Patches (NP) equals the number of patchesof the corresponding patch type (class).
Shape Index:sum of patches’ perimeter divided by the square root of the area of the patch (ha) for each class (class surface) or the entire patch (land surface). This index iscalculated for circle standard (polygon), or square standard (grid) and divided by the number of patches.
Largest Patch Index (LPI) which measures the area of the largest patch in a class divided by the total area of the landscape (multiplied by 100 to convert to percent)
Mean Patch Size (MPS) which measures the average size of a patchin the landscape.
Results and Discussion
District 3 ranked highest and district 1 ranked lowest in ENN indexindicating that urban green space patches in this district were closer together, while green space patches in the third district were limited and far apart from each other. Regarding LPI index, the second district ranked thehighest and the third district ranked the lowest indicating that the largest urban parks in this districtwere much smaller than other districts. Other district had a relatively acceptable statusin this respect. In MPS index, district 2 with 697 patches ranked highest and district 1 with 564 patches ranked lowest indicating that average green space patches in district 1 were smaller. This was also confirmed by maps prepared based on other metrics.Regarding the LSI index, district 1 ranked highest and district 2 ranked lowest, while districts 3 and 4 had a similar status in this measure. The first district had the highest number of patches (NP), while the third district had the lowest NP. The highestPARA ratio was observed in District 1, and the lowestin District 4, while districts 3 and 2 ranked near the middle. In Landscape shape index which increases with the heterogeneity of patches,district 1 (with 13.12) ranked highest and District 3(6.64) ranked lowestwhiledistricts 2 and 4 ranked near the middle.This indicates the heterogeneous shape of green space patches in district 1, while showing that patches of green space in district 3 are very simple and homogeneous.Finally it should be noted that calculating landscape metrics for the four districts ofKhomein indicated a very low per capita green space in this city and also absence of a proper and equitable spatial distribution.
Calculatinglandscape metrics in the four districts of Khomeinindicated thatcompared to other districts, district 1, located in the southern part of the city, has a more desirable status in indices such as PARA, LSI, NP, and ENN. At the same time, district 3, located in the southeastern part of the city, has the least appropriate status regarding these metrics indicating the necessity of a comprehensive analysis of green space areas in this district in near future. Urban managers and planners need to focus on this district and its green space, and if possible find appropriate sites for future green space areas in this district.Although the status of districts 2 and 4, located in the west and north of the cityrespectively, were not very desirable, theyranked higher than districts 3in NP, LPI, and MPS. Using GIS in combination with satellite imagery, and land use metrics provided an innovative way to study the gradual spatial changes in urban green space. Results of landscape metrics analysis indicated an unbalanced distribution of land use in the four urban districts in this study.