Optimizing of landuse using multilayer perceptron neural network in Hamedan city

Subject Areas : Natural resources and environmental management

Naser Shafieisabet ¹ ( Associate Professor, Department of Human Geography and Spatial Planning, Faculty of Earth Sciences, University of Shaheed Beheshti, Tehran, Iran )
Faranak Fezybabaei Cheshmeh Sefeidi ² ( MSc. Student, Department of Geography and Rural Planning, Faculty of Earth Sciences, University of Shaheed Beheshti, Tehran, Iran )

Received: 2021-08-06 Accepted : 2021-10-19 Published : 2023-01-21

Keywords: Neural network, optimizing, land use, Hamedan City,

Abstract :

Background and Objective The protection of natural resources, especially land, and their use, has long been considered. It can be said that because land use has undergone many changes over time, these changes have direct and many effects on the ecosystem and the environment and consequently have various consequences, including these consequences that can be used to change land use. The area was affected by the rapid expansion of urbanization and its effects on land-use patterns in the surrounding environment and, finally, land fragmentation in these areas. Accordingly, in many cases, converting land use from its natural state to artificial land use has irreversible consequences. To reduce the consequences, this can be adapted to the land use structure. Land appropriateness refers to matching the capacities of a plot of land and its land use, and the disproportionate allocation of land use and disregard for its changes has many consequences such as socio-economic segregation, environmental depletion, and loss of resources. Decisions in land and resource management should always be guided in a way that does not conflict with the interests of society and the natural environment. In this regard, one of the effective ways to control and minimize the damage and consequences of land use is to adapt its structure so that, based on the characteristics of land resources and their capabilities, the land can be spatially distributed and arranged more rationally. This study aims to identify and zone the appropriateness of land use structure with the existing capabilities in Hamedan and to evaluate the efficiency of the multilayer perceptron neural network method in the field of land use structure optimization in this city.Materials and Methods In this study, to adapt the land use structure in the city of Hamedan, based on the research background and according to the effective criteria in the field of land use structure, various indicators were selected, including 12 land use indicators, slope, average temperature, and average rainfall. Average humidity, average wind speed, geology, soil type, distance from the river, distance from wells, distance from main roads, and vegetation type. Then, using the field visit, the points with user suitability were registered as educational points. After preparing the layers of the mentioned indicators, these layers were standardized in the software environment of the GIS system. In the next step, the multilayer perceptron neural network uses the after-release algorithm by importing layers affecting the optimization of the land use structure as input and using the middle layer of distance. From appropriate points in terms of land use structure, this network was implemented with the structure of 1-10-12 to adapt the land use structure in Hamedan. From 35% of the total image pixels, the distance from the agricultural proportions as training points falls into three categories the first part (70%) for network training, the second part (15%) for stopping calculations when the error is increasing, and the third part (15%) was used for network verification. Finally, the final land suitability map was drawn. The resulting layer had a value between 0 and 1 which was divided into five land suitability classes. In the present study, after identifying the factors affecting the land use structure and adapting its structure, and preparing each of them, the mentioned layers were standardized. Then, using the field visit, the points with appropriate use were recorded as educational points. Thus, the land use structure was adjusted by the multilayer perceptron neural network model with 58 replications. The results of the neural network validation and the resulting output layer indicate the high accuracy of the network in fitting the land use structure so that the square root mean values of error (RMSE), and absolute error (MAE). Correlation coefficient (R2) in the implementation process of the network is equal to 0.19, 0.21, and 0.89, respectively, indicating the network's high accuracy in implementing the optimizing process. Completely inappropriately divided, and the results showed that most of the areas covered some somewhat suitable and perfectly suitable lands with 32.62 and 28.13% of the total area, respectively.Results and Discussion In the present study, after identifying the factors affecting the land use structure and adapting its structure, and preparing each of them, the mentioned layers were standardized. Then, using the field visit, the points with appropriate use were recorded as educational points. Thus, the land use structure was adjusted by the model of a multilayer perceptron neural network with 58 replications. The results of the neural network validation and the resulting output layer indicate the high accuracy of the network in fitting the land use structure so that the square root mean values of error (RMSE), and absolute error (MAE). The correlation coefficient (R2) in the implementation process of the network is equal to 0.19, 0.21, and 0.89, respectively, which indicates the network's high accuracy in the implementation of the optimizing process. Completely inappropriately divided, and the results showed that most of the areas covered some somewhat suitable and perfectly suitable lands with 32.62 and 28.13% of the total area, respectively.Conclusion The results of optimizing land use structure in Hamedan show that most of the area is not suitable for agricultural activities in terms of effective factors. In this area, most urban land uses completely barren and uncultivable lands, lands. There are mountainous, rocky, and low-quality pastures, mainly in the western and southwestern areas of Hamedan. Also, in this area, lands that have been relatively suitable in terms of a proportion are quite suitable in terms of 12 factors in the best conditions for agricultural and horticultural activities and are the best place for developing agricultural activities. Thus, to change the land use conditions towards a more appropriate trend while paying attention to integrated urban-rural planning for Hamedan and its surrounding settlements. It is recommended to pay attention to land use planning rural-urban plans and projects because the rapid expansion of Hamedan and its suburban spaces has created numerous challenges in terms of land suitability. In such a way that about 23.1% of the lands are ready to be transformed into unsuitable and completely unsuitable conditions. In addition, 32.62% of land use is subject to change to semi-suitable conditions. Based on what has been said, controlling, supervising, and directing the constructions and preventing the over-horizontal expansion of the city of Hamedan and its surrounding spaces by urban and rural stakeholders (local management) is proposed to rangeland and agricultural lands. The findings of the study also indicated that the highest area of land in this area is related to somewhat suitable and perfectly suitable land and the lowest area belongs to unsuitable and somewhat unsuitable land. Therefore, it can be said that the city of Hamedan is currently in a semi-suitable situation in terms of land suitability, which can have a more favorable trend in the future with proper planning and policies.

References:

Bagheri M, Jelokhani Noaryki M, Bagheri K. 2018. Investigation of the land potential of Kermanshah province for rainfed wheat cultivation using artificial neural network. Journal of RS and GIS for Natural Resources, 8(4): 36-48. http://dorl.net/dor/20.1001.1.26767082.1396.8.4.3.2. (In Persian).

Bajocco S, De Angelis A, Perini L, Ferrara A, Salvati L. 2012. The impact of land use/land cover changes on land degradation dynamics: a Mediterranean case study. Environmental Management, 49: 980-989. https://doi.org/10.1007/s00267-012-9831-8.

Benza M, Weeks JR, Stow DA, Lopez-Carr D, Clarke KC. 2016. A pattern-based definition of urban context using remote sensing and GIS. Remote Sensing of Environment, 183: 250-264. https://doi.org/10.1016/j.rse.2016.06.011.

Bharvand S, Soori S. 2015. Landslide hazard zonation using artificial neural network (Case study: Sepiddasht-Lorestan, Iran). Journal of RS and GIS for Natural Resources, 6(4): 15-31. https://girs.bushehr.iau.ir/article_518870.html?lang=en. (In Persian).

Carvajal F, Crisanto E, Aguilar F, Aguera F, Aguilar M. 2006. Greenhouses detection using an artificial neural network with a very high resolution satellite image. In: International Archives of Photogrammetry, Remote Sensing, and Spatial Information Sciences Vol. XXXVI - Part 2 (ISPRS) Technical Commission II Symposium, Vienna. pp 37-42.

Casellas A. 2009. Barcelona’s urban landscape: The historical making of a tourist product. Journal of Urban History, 35(6): 815-832. https://doi.org/10.1177/0096144209339557.

Dadashpoor H, Alidadi M. 2017. Towards decentralization: Spatial changes of employment and population in Tehran Metropolitan Region, Iran. Applied Geography, 85: 51-61. https://doi.org/10.1016/j.apgeog.2017.05.004.

Du T, Vejre H, Fertner C, Xiang P. 2019. Optimisation of ecological leisure industrial planning based on improved GIS-AHP: A case study in Shapingba District, Chongqing, China. Sustainability, 12(1): 33. https://doi.org/10.3390/su12010033.

El-Nozha E-G. 2009. Change detection using neural network with improvement factor in satellite images. American Journal of Environmental Sciences, 5(6): 706-713.

Honarbakhsh A, Pajoohesh M, Zangiabadi M, Heydari M. 2016. Land Use Optimization Using Combination of Fuzzy Linear Programming and Multi Objective Land Allocation Methods (Case Study: ChelgerdWatershed). Iranian Journal of Ecohydrology, 3(3): 363-377. https://dx.doi.org/10.22059/ije.2016.60025. (In Persian).

Hoover EM, Giarratani F. 2020. An introduction to regional economics. Regional Research Institute, Western Virginia University. 444 pp.

Huang H-C, Hwang R-C, Hsieh J-G. 2002. A new artificial intelligent peak power load forecaster based on non-fixed neural networks. International Journal of Electrical Power & Energy Systems, 24(3): 245-250. https://doi.org/10.1016/S0142-0615(01)00026-6.

Jahangir MH, Reineh SMM, Abolghasemi M. 2019. Spatial predication of flood zonation mapping in Kan River Basin, Iran, using artificial neural network algorithm. Weather and Climate Extremes, 25: 100215. https://doi.org/10.1016/j.wace.2019.100215.

Kc B, Race D. 2020. Outmigration and land-use change: A case study from the middle hills of Nepal. Land, 9(1): 2. https://doi.org/10.3390/land9010002.

Kennedy CM, Hawthorne PL, Miteva DA, Baumgarten L, Sochi K, Matsumoto M, Evans JS, Polasky S, Hamel P, Vieira EM. 2016. Optimizing land use decision-making to sustain Brazilian agricultural profits, biodiversity and ecosystem services. Biological Conservation, 204: 221-230. https://doi.org/10.1016/j.biocon.2016.10.039.

Li Y, Li Y, Westlund H, Liu Y. 2015. Urban–rural transformation in relation to cultivated land conversion in China: Implications for optimizing land use and balanced regional development. Land use policy, 47: 218-224. https://doi.org/10.1016/j.landusepol.2015.04.011.

Ma S, He J, Liu F, Yu Y. 2011. Land-use spatial optimization based on PSO algorithm. Geo-spatial Information Science, 14(1): 54-61. https://doi.org/10.1007/s11806-011-0437-8.

Maier HR, Dandy GC. 2001. Neural network based modelling of environmental variables: a systematic approach. Mathematical and Computer Modelling, 33(6-7): 669-682. https://doi.org/10.1016/S0895-7177(00)00271-5.

Onilude O, Vaz E. 2020. Data analysis of land use change and urban and rural impacts in Lagos state, Nigeria. Data, 5(3): 72. https://doi.org/10.3390/data5030072.

Peng J, Wang Y, Wu J, Yue J, Zhang Y, Li W. 2006. Ecological effects associated with land-use change in China's southwest agricultural landscape. The International Journal of Sustainable Development & World Ecology, 13(4): 315-325. https://doi.org/10.1080/13504500609469683.

Pennington DN, Dalzell B, Nelson E, Mulla D, Taff S, Hawthorne P, Polasky S. 2017. Cost-effective land use planning: optimizing land use and land management patterns to maximize social benefits. Ecological Economics, 139: 75-90. https://doi.org/10.1016/j.ecolecon.2017.04.024.

Santaram SO, Rawat YS, Khoiyangbam RS, Gajananda K, Kuniyal JC, Vishvakarma SCR. 2005. Land use and land cover changes in Jahlma watershed of the Lahaul valley, cold desert region of the northwestern Himalaya, India. Journal of Mountain Science, 2(2): 129-136. https://doi.org/10.1007/BF02918328.

Sharda R. 1994. Neural networks for the MS/OR analyst: An application bibliography. Interfaces, 24(2): 116-130. https://doi.org/10.1287/inte.24.2.116.

Subasi A, Ercelebi E. 2005. Classification of EEG signals using neural network and logistic regression. Computer Methods and Programs in Biomedicine, 78(2): 87-99. https://doi.org/10.1016/j.cmpb.2004.10.009.

Taylor MJ, Aguilar-Støen M, Castellanos E, Moran-Taylor MJ, Gerkin K. 2016. International migration, land use change and the environment in Ixcán, Guatemala. Land Use Policy, 54: 290-301. https://doi.org/10.1016/j.landusepol.2016.02.024.

Toure SI, Stow DA, Clarke K, Weeks J. 2020. Patterns of land cover and land use change within the two major metropolitan areas of Ghana. Geocarto International, 35(2): 209-223. https://doi.org/10.1080/10106049.2018.1516244.

Uuemaa E, Antrop M, Roosaare J, Marja R, Mander Ü. 2009. Landscape metrics and indices: an overview of their use in landscape research. Living Reviews in Landscape Research, 3(1): 1-28. http://dx.doi.org/10.12942/lrlr-2009-1.

Xu E, Zhang H. 2013. Spatially-explicit sensitivity analysis for land suitability evaluation. Applied Geography, 45: 1-9. https://doi.org/10.1016/j.apgeog.2013.08.005.

Yilmaz I. 2009. Landslide susceptibility mapping using frequency ratio, logistic regression, artificial neural networks and their comparison: a case study from Kat landslides (Tokat—Turkey). Computers & Geosciences, 35(6): 1125-1138. https://doi.org/10.1016/j.cageo.2008.08.007.