Saputra, Dopit (2023) LAND USE AND LAND COVER PREDICTION IN SURABAYA METROPOLITAN, INDONESIA. Undergraduate thesis, UNIVERSITAS DIPONEGORO; FAKULTAS TEKNIK.

	Text (Cover) Cover.pdf Download (265kB)
	Text (Chapter I) Ch 1.pdf Download (303kB)
	Text (Chapter 5) Ch 5.pdf Download (58kB)
	Text (References) References.pdf Download (196kB)

Abstract

Countless cities worldwide, in both developed and developing nations, have experienced
urbanization. Cities in industrialized nations have experienced urbanization for several
years. In contrast, cities in emerging nations, particularly in Asia, are at an early stage of
urbanization. Several cities in Indonesia, the fourth most populous nation, have undergone
an early stage of urbanization. Therefore, this study aims to classify cities and their growth
rates in Indonesia. Using the city of Surabaya as an example, the spatial expansion of the
cities has been forecasted. The classification of cities in this study is based on principal
component analysis and cluster analysis. Then, LULC projection will be conducted using
two different methods; MLPNN and ANN-CA. First, The Land Change Modeler from
TerrSet was used to implement the multi-layer perceptron neural network-based Markov
chain model method in order to obtain the land use and land cover (LULC) changes
projection. The second method, ANN-CA, was used to similar data for the same objective.
Afterwards, a comparison of the two approaches is presented. The study's findings indicate
that 35 cities can be divided into five groups. Besides, the LULC changes study in Surabaya
revealed a large expansion of 60 percent between 2000 and 2021, with a severe decline in
agricultural land area of over 20 percent, amounting to 132.62 km
during the same time
frame. Besides, the results of the transition probability matrix indicate that the tendency of
transformation into an urban area is relatively high, with values of 0.0903 and 0.1033 for
vegetation and agricultural, respectively. However, the change from other land cover
classes to urban areas was not as apparent in the shorter time frame of 2000-2003, with a
tendency of change from vegetation of 0.0891 and from agricultural land of 0.0956. The
projection results indicate that the urban area will be just under 350 km
2
and stretch
southward from Gresik (the west side of Surabaya). The main difference between MLPNN
and ANN-CA is how the models are built. MLPNN uses feedforward, while ANN-CA uses
both feedforward and Cellular Automata.

2

Actions (login required)

View Item