SPATIOTEMPORAL PATTERNS AND ITS INSTABILITY OF LAND USE CHANGE IN FIVE CHINESE NODE CITIES OF THE BELT AND ROAD

It has long recognized that there exists three different terrain belt in China, i.e. east, central, and west can have very different impacts on the land use changes. It is therefore better understand how spatiotemporal patterns linked with processes and instability of land use change are evolving in China across different regions. This paper compares trends of the similarities and differences to understand the spatiotemporal characteristics and the linked processes i.e. states, incidents and instability of land use change of 5 Chinese cities which are located in the nodes of The Silk Road in China. The results show that on the whole, the more land transfer times and the more land categories involved changes happens in Quanzhou City, one of eastern China than those in central and western China. Basically, cities in central and western China such as Changsha, Kunming and Urumuqi City become instable while eastern city like Quanzhou City turns to be stable over time. * Corresponding author The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2/W7, 2017 ISPRS Geospatial Week 2017, 18–22 September 2017, Wuhan, China This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-2-W7-1327-2017 | © Authors 2017. CC BY 4.0 License. 1327


INTRODUCTION
Human activities make the range gradually transition from the natural ecosystem to the artificial ecosystem in the area around the city, which gradually become a region of intense human activity.The complex system coupled urban-nature system has a high degree of complexity, for local, regional and even the development of the global economy has a significant role.A complex system with the main elements of the humanities and the natural factors, the exchange of material flow, energy flow and information flow continuously with the external system, this is the result of the interaction between urban and natural complex system and other elements (Costanza, 1997;Liu, 2007;Chen et al., 2014).In 2008, more than 50% of the world's population has been living in urban areas.by 2050, the population will be living in the city of 66%, while the rate of urbanization in China reached 90% (Chneider et al. 2015).Study shows that the global city of the 67 regions of the world in 1970-2000, China is one of the most rapid cities in the process of urbanization (Seto et al., 2011).More than 30 years of reform and development, the population has increased by 500 million, urbanization rate increased from 19.4% to 49.2 %, increased by 29.8 %, as of 2012, more than 50% of China's population lives in urban areas (Chen et al., 2013;Kuang et al., 2014).
China has become the most rapid development of urbanization in the world, which has been widely concerned by domestic and foreign scholars.Gu et al.,(2014) studied the change characteristics of urbanization in China from five stages  and the impact of urbanization on the process of urbanization.Mertes et al., (2015) using MODIS data to study the urban land use change of the thirteen countries in East.By contrast, China is the most rapid development of urbanization in East Asian countries.Chi et al., (2015) compared with China and the United States 1978-2010 urban internal space and time differences, analysis of the differences in urban land use structure of different countries, and to explore the reasons for the differences.Huang et al., (2015) studied the changes of urban land use in the three regions of the eastern, central and western regions.Liu et al., (2015) further comparative analyzed the changes of urban land use in the East, northeast, central, western regions and Hong Kong, Macao and Taiwan regions in five regions of urban expansion.
With almost 30 years reformation and opening to abroad, China's urban economic system has gradually formed, the different scale, levels and functions of the city has a diverse range of links.Wu et al., (2015)use DMSP-OLS NIGHTTIME LIGHT (NTL) data to verify the levels of city.And the Chinese city is divided into seven urban nodes, 26 regional urban nodes, 107 nodes provincial city, which is basically the same division status as that of Gu did (Gu, 2008).Liu et al., (2012) regionalize China into 5 regions and analyzed the changes of land use during nearly two decades.In addition, Huang et al., (2015) applied the "first classification in grade, and second in class" ideas and Q cluster method to analyze regional difference and 23 Chinese urban structures were divided into 4 levels, i.e. nation, region, sub-region and locality.
These studies are mainly based on the regional level,

Data sources
Remote sensing images and their resolution, data sources, data format of Quanzhou, Changsha, Yinchuan, Urumuqi, Kunming are shown in Table 1 which is in accord with the 1: 100,000 land use classification.We consulted the land cover  1989,1995,2000,2007,2014

Incidents and States Algorithms
In order to study the processing of urban land use Table 2.The Flow Matrix form (Runfola, Pontius, 2013) (2) Based on Flow Matrix to calculate urban land use change instability Eq. 1 and Eq. 2 are the foundation of calculating urban land use instability.Eq.1 changes calculated observe intensity S in each time interval; the Eq.2 calculated uniform intensity U of the entire study period.After calculating the observe intensity in each time interval, we compared with uniform intensity which is compares S to U. if S > U, then the change is relatively fast for that time interval.If S <U, then the change is relatively slow for that time interval.(Aldwaik, Pontius, 2012;Pontius, Gao et al., 2013； Zara, Joã o, Pontius, 2016;Enaruvbe, Pontius, 2015).
The maximum value of R can be manifeste  1990,1995,2000,2005,2010100 Changsha Cropland, Grassland, Forest 1990,1995,2000,2005,2010100 Kunming Forest, Grassland, Built 1990,2000,2008,2014100 Yinchuan Cropland, Built, Grassland 1989,1995,2000,2007,201430 Urumuqi Cropland, Grassland, Built, Unused land 1990,1999,2006,2014 100 Tab and these findings provided much insight into land use change patterns, especially in large cities and large regions.However, a comprehensive examination of these findings with medium cities in the three different terrain of China is lacking up to date.It remains unclear that the differences of spatiotemporal patterns and processes such as states, incidents and instability of cities in different regions of China.To address these questions requires selecting and comparing several cities together using consistent method and data.Changsha, Kunming, Yinchuan and Urumqi and Quanzhou City are selected for this study due to the similarity in urban size and economy level.Moreover, they are respectively located in the eastern, central and western China, which is helpful to compare the differences in land change pattern and process (Figure.1).In addition, they are major node cities on the Belt and Road and located in the Silk Road Economic Belt and the 21st-Century Maritime Silk Road.Thus they have important traffic location and the five cities form the "East -Central -West" transition.In this study, the primary research objective The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2/W7, 2017 is to identify the similarities and differences in the spatiotemporal characteristics and the linked processes i.e. states, incidents and instability of land use change of 5 Chinese cities.

Figure 1 .
Figure 1.Location map of study area in the Belt and Road of China changes, we applied a computer program in the language Visual Basic for Applications embedded in Microsoft Excel that interfaces with the GIS software TerrSet (Eastman 2014).A pixel's number of Incidents is the number of times the pixel experiences a change.Incidents can range from 0, indicating complete persistence, to the number of time points minus 1, indicating change across all time intervals.A pixel's number of States is the number of different categories that the pixel represents at all time points.States can range from 1, indicating complete persistence, to the smaller of the number of time points and the number of categories(Zhang, 2011;Runfola, Pontius, 2013;Zhang , Pontius, 2016).The two algorithms measure the multi-phase remote sensing image classification in spatial in different angles.The more detail is shown in Figure2.

Figure 2 .
Figure 2. The number of land use the Incidents and the States algorithms Eq.3 calculated land category change instability index R in differently intervals.Numerator is the maximum of the observe intensity in special time interval and uniform intensity in different time interval.And the denominator is uniform intensity multiplying by the whole time interval.The ratio of numerator and denominator is R, which is a calculated measure of urban land use change.If R=0, the change is absolute stability, that is S = U.The R >0, the change is instable.

dFigure 3 .
Figure 3. provide maps of each city in which legend colors indicate more frequent transitions over . 3 The characteristics of typical city land use data incidents Map of Quanzhou is shown in Figure 3(a).It indicated that the frequency mainly dominated by 1 time and 2 times, which are accounted for 31.5% and 41.4%, respectively.These changes are mainly located in the ring of Quanzhou Bay Coastal and economic development Zone of Quanzhou.It is because that it is flat for central and southeast of Quanzhou, and so human activities often put a pressure on the land use.It is accounted for 18.5% that the Transition Frequency of times reaches three times.These changes occurs mainly in Dehua,Yongchun and Jinmen County.There is a cycle of fast-growing of forest cultivation, so the remote sensing images show a "Forest land -woodlandorchard -artificial turf -orchard -woodland "change sequence(Zhang et al, 2012).

Figure 3 .
Figure 3.The Incidents map in the typical cities.

Figure 3
Figure 3(b) is a four interpretation of remote sensing data obtained the transition frequency of Urumuqi urban land use.The 1 time land category transition is mainly accounted for the 91%.Generally, Cropland, Grassland and Bare were converted to the Built, which happens mainly in the Midong and Tianshan District.The 2 times only accounted for 8.6% whose land conversion is as follows: Grassland → Cropland →Built.Figure 3 (c)-(e) shows the times of transition for Kunming City.In Kunming, the 1 and 2 times of land transition are accounted for 34.7% and 64.8%, respectively.In the northwest of Luquan Yi and Miao Autonomous County and Dongchuan District are mainly forest, Cropland and Grassland.Panlong, Wuhua Guandu District is mainly construction land.It appears around the Xishan, Jinning, Chenggong District are water in the form of the distribution network.The 1 time and 2 times transfer are mainly for Cropland, Forest and Grassland

Figure 4
Figure4shows the comparison of land transfer frequency of five typical cities. Kunming and Urumuqi both mainly occur to divert 1 time,

Figure 4 .
Figure 4.The land use change of Incidents We calculate the observed changes S and uniform intensity based on Eq. 1 and Eq. 2 in Figure 5.We can see that except the Yinchuan City, the cities in central and western China such as Changsha, Kunming and Urumuqi City become instable while eastern city like Quanzhou City turns to stable over time.Obviously, from the instability perspective, the

Figure 5 .
Figure 5. Land use instability: (a) The most stable cities, (b) The mid-ranked instability, (c) -(e)The most instability

Figure. 6
Figure.6 Comparison of the land use change with instability in typical cities.

Figure. 8
Figure.8 compares the regional differences and commonalities of land use change states among the five typical cities. Quanzhou, Changsha, Yinchuan and Urumuqi mainly had the transfers between 2 categories, which accounts for their 75%, 96%, 85.9% and 95.4%, respectively.Also, Kunming and Quanzhou Cities have a longer bar than the other cities in the "3 to 5 categories transfers".But

Figure. 8
Figure. 8 The States changes in land use change

Table 1
. The study area data sources2.2Methods