Based on the collected data, we first used the AHP method to evaluate UDTI, PTI, LTI, STI and ITI in 1996 and 2012, and the changes in the indexes during the 16 years around the Bohai Rim Region. On the basis of the calculation from the AHP method, we then used the ESDA method to calculate the global Moran’s I index and Local Moran’s I index to investigate each indexes’ spatial agglomerate rules and characteristics, and to draw LISA cluster maps for each index to study the significance level of spatial agglomeration between one regional unit and the other.
Based on the urban-rural development transformation evaluation system that had previously been constructed, we calculated the PTI, LTI, ITI and STI for 1996 and 2012, and the change in the indexes during the 16-year period.
3.1.1 The spatial-temporal pattern of PTI
PTI shows the urbanization trend, the proportion of the labor force, the average level of education, and the level of aging in urban and rural regions. As shown in
Table 2, the average PTI around the Bohai Rim Region increased by 0.11, with the PTI for Beijing increasing the fastest, from 0.48 to 0.70, and the PTI for Liaoning increasing the slowest, from 0.34 to 0.47. Using the natural break classification method in ArcGIS we divided the PTI results into five ranges: low value, relatively low value, medium value, relatively high value, and high value. In 1996 the counties were mainly in the relatively low value and low value ranges, which accounted for 36.0% and 27.4% of the total number of counties respectively. The numbers of counties in the relatively high value and high value ranges were comparatively low, at 31 (9.5%) and 26 (7.9%) respectively. In 2012 the classification results had changed greatly. Although the relatively low value range still contained the most counties, the number had dropped from 118 to 107, with the percentage points decreasing from 36.0% to 32.6%. The range with the second highest number changed from the low value to the medium value: the number of counties in the medium value range increased from 68 to 95, while the number in the low value range fell from 90 to 46. The high value and low value ranges became the two smallest ranges, accounting for 9.8% and 14% of the counties respectively.
In order to investigate the extent of PTI change, we used ArcGIS to plot the results on a map. The calculation showed that from 1996 to 2012, the fastest population evolution areas were located mainly in the municipalities of Beijing, Tianjin, and in some of the counties in Langfang, Shijiazhuang, Handan and Chaoyang cities; the areas evolving relatively rapidly included counties in Beijing and Tianjin, southern and central-eastern parts of Hebei, eastern and southern parts of Shandong such as Qingdao, Yantai, and the central part of Liaoning. The medium range areas were located mainly in the southeastern and northern parts of Hebei, western and central parts of Liaoning, and southern and central parts of Shandong such as Linqu. The relatively slow areas included the northwestern part of Hebei, the eastern part of Shandong, and most cities in Liaoning. The slow range areas were located mainly in the northwestern part of Liaoning (
Figure 1).
Figure 1 PTI change from 1996 to 2012 |
3.1.2 The spatial-temporal pattern of LTI
LTI, which reflects the overall trends of the proportion of urban and rural construction land, the proportion of arable land, and the proportion of countryside homestead, can be used to evaluate the land-use structure and land occupancy that includes housing constructions for urban and rural residents.
According to our calculations, the average LTI increased from 0.32 to 0.34 between 1996 and 2012. Tianjin experienced the fastest increase, from 0.37 to 0.47; Liaoning increased relatively slowly, from 0.41 to 0.43; Hebei increased the slowest, from 0.30 to 0.31, a rise of only one percentage point during the 16-year period.
The LTI result was also divided into five ranges (low value, relatively low value, medium value, relatively high value, and high value). In 1996, the relatively low value range contained the most counties, at 29.9%; the range with the second highest number was the medium value range, containing 82 counties, at 25.1%; the high value and relatively high value ranges were the two smallest classifications, accounting for only 7.3% and 17.1% of all the counties respectively. By 2012 the number of counties in the medium value range had increased rapidly, from 82 to 108. The range with the second highest number was the relatively low value range, which had decreased by 2.5 percentage points during the 16 years.
The low value and relatively high value ranges also experienced a rapid decrease. Taking the relatively high value range as an example, during the 16-year period the number of counties in this range decreased from 56 to 46. The high value range still contained the fewest counties, increasing by only one county.
We used ArcGIS to plot the LTI change results on a map. The municipalities of Beijing and Tianjin and the cities of Jinan, Zibo and Qingdao experienced rapid increases during the 16-year period. The relatively fast evolution areas were located mainly in counties of Beijing and Tianjin, eastern parts of Hebei such as Tangshan, central and northern parts of Shandong such as Binzhou, and central and southern parts of Liaoning such as Dalian. The medium and relatively slow evolution areas covered most parts of the three provinces. The slow evolution areas were located mainly in the southern part of Hebei and the northern part of Liaoning (
Figure 2). ITI is an integrated measurement of the level of regional non-agricultural industrial development and agricultural mechanization.
Figure 2 LTI change from 1996 to 2012 |
3.1.3 The spatial-temporal pattern of ITI
The average ITI increased rapidly, from 0.10 to 0.48 during the 16-year period. In the five provincial administrative units, Shandong increased the fastest, from 0.09 to 0.49, and Hebei increased relatively fast, from 0.09 to 0.45.
Based on the classification of ITI, we realized that a great change had taken place during the 16-year period. In 1996, the low value range contained the most counties of all, 103, including the southern part of Hebei and the western part of Shandong. The relatively low value contained the second highest number of counties, 97, accounting for 29.6% of the total. The relatively high value range contained the fewest, with 8.5%. In 2012, the number of counties in the low value and relatively low value ranges had decreased, especially the low value range, where the number decreased from 103 to 48, while the counties in the medium value, relatively high value and high value ranges increased, especially the relatively high value range, where the number rose from 28 to 75.
Figure 3 ITI change from 1996 to 2012 |
The results on the map (
Figure 3) show that the areas of rapid evolution were located mainly in administered districts of Liaoning such as Fushun and Panjin cities. The relatively fast areas were in the counties of Beijing and Tianjin, eastern and southern parts of Hebei such as Tangshan and Handan cities, and central and eastern parts of Shandong such as Tai’an, Laiwu and Yantai cities. The medium and relatively slow evolution areas covered most of the three provinces and some of the districts in Beijing and Tianjin municipalities. The slow evolution areas were located mainly in Liaoning, such as Fuxin and Fengcheng cities.
3.1.4 The spatial-temporal pattern of STI
The STI shows the regional socio-economic development, the income of urban and rural residents, the public service and social security conditions, as well as the maturity and quality of local residents’ social environment and the gap in the living standards between urban residents and rural residents.
According to the evaluation, the average STI in the Bohai Rim Region had risen from 0.14 to 0.24 during the 16-year period. Shandong increased the fastest among the STIs for the five provincial administrative units, from 0.14 to 0.27. Beijing and Liaoning increased relatively fast, increasing by 15 and 12 percentage points respectively. The STI of Tianjin increased relatively slowly, from 0.18 to 0.29. Hebei’s STI increased the slowest, from 0.13 to 0.20.
With regard to the classification of STI, in 1996 the medium value range contained the most counties of all, 130, accounting for 39.6%; the relatively high value range contained fewer, 96, followed by the relatively low value range, 64 counties, the low value range, 23 counties, and the high value range, which contained the fewest counties, 20, accounting for only 6.1% of the total. In 2012, the relatively low value range became the classification with the most counties, 117, accounting for 35.7% of the total; the range with the second highest number was the medium value range, containing 94 counties; the relatively high value range decreased the fastest, from 96 to 59 counties; the low value range experienced an increase, from 23 to 42 counties, with the percentage rising from 7.0% to 12.8%; the high value range still contained the fewest counties, at only 21.
The map shows that Beijing and the cities in the eastern part of Shandong province such as Qingdao, together with the southern part of Liaoning province, namely Dalian and Rongcheng experienced a rapid increase in social evolution. The eastern part of Hebei, the northern part of Shandong and the central part of Liaoning increased relatively quickly. The areas of slowest evolution were located mainly in the northern and western parts of Hebei, such as Zhangjiakou City (
Figure 4).
Figure 4 STI change from 1996 to 2012 |
3.1.5 The spatial-temporal pattern of UDTI
UDTI makes an assessment of the overall pattern of urban-rural development transformation in the Bohai Rim Region. It can be calculated using the following formula:
The average UDTI of the Bohai Rim Region increased from 0.23 to 0.38 during the 16-year period. Among the five provincial administrative units, the UDTI of Tianjin increased the fastest, from 0.27 to 0.47; the index increased relatively fast in Shandong and Hebei, rising by 16 and 15 percentage points respectively; Liaoning’s UDTI increased the slowest, from 0.29 to 0.42.
We also classified the calculation results into five ranges. In 1996, the highest number of counties was in the relatively low value range, with 114 counties, accounting for 54.7%; the range with the second highest number was the low value range, followed by the medium range and relatively high value range; the high value range contained the fewest counties, 16 counties, accounting for only 4.9% of the total. In 2012 the number of counties in the medium value and high value ranges increased, while the others decreased. The high value range increased the fastest, from 16 to 32 counties; the relatively low value range decreased the fastest, from 114 to 100 counties.
The results of the classification of the change in UDTI during the 16-year period showed that the areas of fastest evolution were in Beijing, Tianjin, and big cities from the three provinces. The relatively fast evolution areas were mainly in districts of Beijing, Tianjin, the southern part of Hebei, such as Handan City, and central and eastern parts of Shandong, such as Laiwu City, Dongying City and Yantai City. The medium and relatively slow areas of evolution were in most parts of Hebei and Liaoning, and western and southern parts of Shandong. The slowest evolution areas were located mainly in Liaoning, such as Fuxin City and Chaoyang City (
Figure 5).
Figure 5 UDTI change from 1996 to 2012 |
The evaluation of the UDTI, PTI, LTI, ITI and STI using the AHP method around the Bohai Rim Region in 1996 and 2012 enabled us to ascertain the urban and rural development and transformation process around the region. In this section, we use the ESDA method to detect the spatial distribution of index change during the 16-year period.
3.2.1 Global Moran’s I index for each index
We found that all of the indexes except the ITI presented positive spatial autocorrelation under the 0.05 significance level test. The STI change presented the highest Moran’s I index, 0.33, which meant that social transformation had the most significant spatial autocorrelation level; the ITI change had the lowest Moran’s I index, with the
Z value test below 1.96, showing no significant spatial autocorrelation characteristics. The Global Moran’s I index of UDTI change was 0.13, which is relatively low compared with other indexes, implying a weak positive correlation in spatial terms (
Table 2).
Table 2 Global Moran’s I index of the five indexes around Bohai Rim |
| Index | UDTI change | PTI change | LTI change | ITI change | STI change |
| Global Moran’s I | 0.13 | 0.31 | 0.27 | 0.04 | 0.33 |
| Z value | 5.20 | 12.27 | 10.88 | 1.59 | 13.10 |
3.2.2 Local Moran’s I index for each index
The Global Moran’s I index reflected the integral spatial distribution of each index. We used the Local Moran’s I index to detect specific clustered districts of high values and low values at the local level, in order to identify the location of spatial heterogeneity.
Figure 6 shows the calculation results for the Local Moran’s I index of PTI, LTI, ITI, STI and UDTI change under the 0.05 significance level test. The PTI change reflected the fact that the spatial distributions of most indexes were in the High-High (H-H) and Low-Low (L-L) clusters. The H-H clusters were concentrated mainly in Beijing, Tianjin, and the southern part of Hebei, such as Shijiazhuang City; the L-L clusters were concentrated in the northern part of Liaoning, such as Fushun City. The LTI change reflected the fact that the H-H clusters were concentrated mainly in Beijing, Tianjin, and the districts of Dongying and Qingdao cities in Shandong, while the L-L clusters were concentrated mainly in Handan, Xingtai and Hengshui cities in the southern part of Hebei. The spatial distribution of ITI change was not significant around the Bohai Rim. However, the relatively high value cluster was in Qingdao City. The H-H clusters for STI change were located mostly in Shandong, such as Dongying, Zibo, Qingdao and Yantai cities, and the southern part of Liaoning, such as Dalian and Anshan cities. The L-L clusters were located in southern and western parts of Hebei.
Figure 6 The calculation of Local Moran's I index for UDTI and four indexes |
The spatial distribution of UDTI change had features in common with other indexes. Beijing, Tianjin municipality and Qingdao City remained in H-H clusters, which meant that they had experienced a rapid urban and rural development and transformation process during the 16-year period. The L-L clusters were mainly in areas such as Fuxin and Chaoyang cities in Liaoning, and Hengshui City in Hebei, representing a low speed of evolution for urban and rural development and transformation. Using the index maps we identified two zones that consistently showed a high speed of development and transformation, namely the Beijing-Tianjin circle and the Qingdao circle. The reason for the formation of these two circles will be discussed in the following section.
