人穷还是地穷?空间贫困陷阱的地统计学检验
作者简介:马振邦(1983- ),男,甘肃会宁人,讲师,研究方向为景观地理与区域可持续发展。E-mail: zbma@lzu.edu.cn
收稿日期: 2018-04-16
网络出版日期: 2018-10-20
基金资助
国家自然科学基金项目(41401204,41471462)
中央高校基本科研业务费项目(lzujbky-2013-128)
Poor people, or poor area? A geostatistical test for spatial poverty traps
Received date: 2018-04-16
Online published: 2018-10-20
Copyright
马振邦 , 陈兴鹏 , 贾卓 , 吕鹏 . 人穷还是地穷?空间贫困陷阱的地统计学检验[J]. 地理研究, 2018 , 37(10) : 1997 -2010 . DOI: 10.11821/dlyj201810010
The test for spatial poverty traps (SPTs) is a hot issue in the field of the geography of rural poverty. However, the main existing approaches cannot provide spatial scale-related information, which may be a restriction on gaining a deeper understanding of the mechanism of SPTs. Therefore, we conducted a case study in the Liupan Mountain Region by introducing geostatistical methods. The semivariogram and cross-correlogram were employed to quantitatively describe the spatial pattern of village-level poverty and its relationship with the selected geographical factors respectively, so that the scale-dependent spatial form and underlying reasons for SPTs can be explored. The village-level poor population (PP) and poverty rate (PR) were used as the poverty indicators. The results show that the geostatistical methods can provide satisfactory and reliable performance in the test for SPTs: (1) The semivariogram models can indicate both the spatial structure and the autocorrelation range of the two indicators, which can describe the extent and the range of the spatial form of SPTs (i.e. the spatial aggregation of poverty). The percentages of the random variance (nugget, C0) in the total variance (sill, C0 + C) are 34.4% and 11.5% for PP and PR, respectively. The range of autocorrelation is 9.3 km for PR, and 5 and 48 km for PP. (2) The cross-correlograms further show that the two indicators are significantly (P<0.05) correlated with the geographical factors within different spatial ranges. Generally, the poverty status of a village is mainly in response to three factors (i.e. the distance to the nearest county town, the elevation, and the total population) within a wide range. In conclusion, the evidence of SPTs from our work is consistent with the reality that the study area has suffered persistent poverty in the past three decades.
Fig. 1 Location of the study area and spatial distribution of the villages, the rivers and the roads图1 案例区位置及行政村、水系、道路分布图 |
Fig. 2 Frequency distribution histograms for the poverty indicators图2 贫困指标的直方图 |
Tab. 1 Description and type of the geographical indicators at the village level表1 村级地理指标描述及类型 |
| 地理指标 | 指标描述及含义 | 类型 | 预期 |
|---|---|---|---|
| 平均海拔 | 村域海拔均值,表征地形特征及其关联的温度等农业生产条件 | N | + |
| 平均坡度 | 村域坡度均值,表征地形特征及其关联的土壤、侵蚀等农业生产条件 | N | + |
| 年降雨量 | 村域年降水量均值,表征水分等农业生产条件 | N | - |
| 总人口数 | 村总人口数,表征人口及社会关系状况 | H/S | +/- |
| 到河流距离 | 村点到最近主要河流距离,表征土壤及可灌溉程度等农业生产条件 | P | + |
| 到县城距离 | 村点到最近县区政府驻地时间,表征较高水平教育、医疗、市场、信息等服务和非农就业机会的可达性程度,以及较强非农社会关系状况 | P/S | + |
| 到乡镇距离 | 村点到最近乡镇政府所在地时间,表征教育、医疗、市场、信息等服务和非农就业机会的可达性程度,以及非农社会关系状况 | P/S | + |
| 到道路距离 | 村点到最近县级及以上道路时间,表征交通基础设施条件 | P | + |
注:N/P/H/S为地理资本类型:N为自然资本;P为物质资本;H为人力资本;S为社会资本。 |
Tab. 2 Descriptive statistics of the indicators at the village level表2 村级贫困指标及地理因子的描述性统计 |
| 具体指标 | 平均值 | 中位数 | 标准差 | 偏度 | 峰度 | 最小值 | 最大值 | |
|---|---|---|---|---|---|---|---|---|
| 贫困状况 | 贫困人口数(人) | 445.1 | 388 | 307.4 | 1.43 | 3.28 | 24 | 2248 |
| 贫困发生率(%) | 33.1 | 33.4 | 17.5 | 0.26 | -0.085 | 1.2 | 87.7 | |
| 贫困人口数a | 18.69 | 18.88 | 5.26 | 0.04 | -0.21 | 5.65 | 36.30 | |
| 贫困发生率b | -0.75 | -0.73 | 0.22 | -0.03 | -0.35 | -1.21 | -0.12 | |
| 地理因子 | 平均海拔(m) | 1916 | 1961 | 281 | -0.33 | -0.21 | 1225 | 2826 |
| 平均坡度(°) | 12.21 | 12.61 | 3.50 | -0.61 | 1.32 | 1.26 | 23.69 | |
| 年降雨量(mm) | 459.7 | 463.4 | 31.6 | 0.35 | -0.26 | 396.9 | 546.1 | |
| 总人口数(人) | 1441.1 | 1246 | 783.7 | 2 | 6.16 | 263 | 6808 | |
| 到河流距离(km) | 3.89 | 3.30 | 3.07 | 0.71 | -0.29 | 0.02 | 14.15 | |
| 到县城距离c(min) | 47.78 | 44.00 | 27.53 | 0.99 | 2.04 | 1.00 | 236.00 | |
| 到乡镇中心距离c(min) | 20.05 | 14.00 | 19.76 | 2.12 | 5.31 | 0.00 | 114.00 | |
| 到县级以上道路距离c(min) | 5.30 | 0.00 | 9.78 | 2.51 | 6.93 | 0.00 | 67.20 |
注:a表示Box-Cox正态转化后结果(λ=1/3);b表示Box-Cox正态转化后结果(λ=0.8);c距离表示车行时间。 |
Fig. 3 Semivariogram curves of the poverty indicators图3 贫困指标的变异函数曲线 |
Tab. 3 Isotropic semivariogram models and parameters of the poverty indicators表3 各向同性条件下贫困指标变异函数的理论模型及参数 |
| 贫困指标 | 理论模型 | 块金值 | 偏基台值 | 变程(km) | 基台值 | 块金值/基台值(%) | 决定系数 |
|---|---|---|---|---|---|---|---|
| C0 | C | a | C0+C | C0/(C0+C) | R2 | ||
| 贫困人口数 | 指数模型 | 14.3 | 14.3 | 31.8 | 28.6 | 50 | 0.926 |
| 套合模型:球状I | 10.3 | 11.1 | 5 | 29.9 | 34.4 | 0.984 | |
| 球状II | 8.5 | 48 | |||||
| 贫困发生率 | 指数模型 | 0.0058 | 0.0446 | 9.3 | 0.0504 | 11.5 | 0.968 |
Fig. 4 Cross-correlograms between the poverty indicators and the geographical factors图4 贫困指标与地理因子的交叉相关图 |
Fig. 5 Values of AR between the poverty indicators and the geographical factors图5 贫困指标与各地理因子之间的平均空间相关系数 |
Fig. 6 Cross-correlograms between the poverty rate and the geographical factors at different lag intervals图6 不同步长下贫困发生率与地理因子的交叉相关图 |
Tab. 4 Isotropic semivariogram models and parameters of the poverty rate in different sub-regions表4 次区域贫困发生率变异函数的理论模型及参数 |
| 次区域 | 贫困发生率 均值(%) | 理论模型 | 块金值 | 偏基台值 | 变程(km) | 基台值 | 块金值/基台值(%) | 决定系数 |
|---|---|---|---|---|---|---|---|---|
| C0 | C | a | C0+C | C0/(C0+C) | R2 | |||
| 分区1 | 29.7 | 指数模型 | 0.0021 | 0.0146 | 19.1 | 0.0167 | 12.6 | 0.88 |
| 分区2 | 35.1 | 指数模型 | 0.0054 | 0.0324 | 6.3 | 0.0378 | 14.3 | 0.36 |
| 分区3 | 37.5 | 指数模型 | 0.0119 | 0.0693 | 9.7 | 0.0812 | 14.7 | 0.96 |
| 分区4 | 31.5 | 指数模型 | 0.0138 | 0.0420 | 15.0 | 0.0558 | 24.7 | 0.95 |
Fig. 7 Cross-correlograms between the poverty rate and the geographical factors in different sub-regions图7 次区域贫困发生率与地理因子的交叉相关图 |
The authors have declared that no competing interests exist.
| [1] |
|
| [2] |
|
| [3] |
[
|
| [4] |
[
|
| [5] |
[
|
| [6] |
[
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
[
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
[
|
| [17] |
|
| [18] |
[
|
| [19] |
[
|
| [20] |
|
| [21] |
|
| [22] |
[
|
| [23] |
|
| [24] |
[
|
| [25] |
[
|
| [26] |
|
| [27] |
[
|
| [28] |
|
| [29] |
|
| [30] |
[
|
| [31] |
[
|
| [32] |
[
|
| [33] |
[
|
| [34] |
|
| [35] |
|
| [36] |
[
|
/
| 〈 |
|
〉 |