基于多特征组合优选与随机森林算法的石漠化信息提取——以云南省昭通市为例

石漠化是我国西南地区最主要的地质生态灾害之一,其引发的土地资源丧失、生态系统退化、干旱缺水等问题严重威胁着地区的生态安全、粮食安全和不发生规模性返贫。因此精确提取石漠化信息对区域经济社会持续发展至关重要。针对当前石漠化信息提取中存在的时相单一、时效性差、区域尺度提取结果精度低等问题,以云南省昭通市为例,提出一种多特征组合优选的分类方法。在优选样本和特征的基础上,利用Sentinel-2影像和DEM数据提取光谱、指数、植被覆盖度、基岩裸露率、纹理、地形等多特征,构建5种分类方案,并采用随机森林分类算法完成提取。结果表明:2020年昭通市石漠化土地面积为2820 km2,占全市土地利用/覆被面积的11.11%,分类结果与实地调查区域一致性较好;利用Jeffries-Matusita distance(JM距离)得到的特征优选方案总体精度为88.0%,Kappa系数为0.85,石漠化土地生产者精度和用户精度分别达到91.2%和83.8%。本研究提出的方法能够较为准确地获取区域尺度石漠化空间分布信息,可为相关部门开展石漠化防治与监测工作提供参考。

Effects of Enclosure on Grassland Ecological and Economic Benefits in Northern China

Grasslands in northern China serve the country as both an ecological barrier and a livestock production base.There,installing enclosures has been becoming the major grassland restoration measure adopted by many local governments.However,the effects of restoration on both ecological and production benefits of grassland remain unclear for implemented grassland restoration policies.Therefore,a representative rangeland in northern China,the Maodeng pasture in Inner Mongolia Autonomous Region was selected as the study area,and remote sensing monitoring analyses were carried out to quantify the ecological benefits and economic benefits from 2015 to 2021.The results showed that:1) in terms of ecological benefits,the grassland area with a grassland coverage rate of more than 60% accounts for 32.3% of the regional area,and 86.4% of its grassland grew significantly better than the same period in2015,showing a significant improvement in grassland growth.Using the average amount of carbon per unit area as the ecological benefit evaluation index,it increased by 27.1% to 32.48Tg C/yr from 2015 to 2021.2) In terms of economic benefits,both theoretical grass production and livestock carrying capacity increased from 2015 to 2021.Compared to 2015,the theoretical grass production in 2021 increased by 24.8% to 71 900 t.The livestock carrying capacity reached 52 100 sheep units in 2021,nearly 11 000 sheep units more than that in 2015.During the study period,multiple economic indicators(on a per capita basis of permanent residents) for the pastoral area of Xilinhot City to which the Maodeng pasture belongs,have grown steadily.Per capita total income rose from 29 630 yuan(RMB) in2015 to 62 859 yuan(RMB) in 2021.Relying on grassland resources to develop the pastoral ecology also broadens the potential economic development space.Overall,the establishment of the reserve and the experiment of implanting an enclosure policy have had a significant and positive impact on Maodeng pasture’s development from both an ecological and economic perspective.With the support of scientific evidence,enclosure policy can be extended to more than 110 000 km~2 of grasslands in northern China with similar precipitation and temperature conditions,enhancing the productive and ecological potential of grasslands.The above research results will contribute to the scientific formulation of grassland pasture quality improvement plans in northern China.

基于GF-6 WFV影像和CSLE模型的山区耕地侵蚀定量评价及特征分析

坡耕地是高原山区水土流失最主要的策源地,其严重的土壤侵蚀已经威胁到山区粮食安全和农业可持续发展。该研究以云南省土壤侵蚀较为严重的文山州为例,在第三次全国国土调查数据、水利普查抽样单元和GF-6WFV等多源遥感数据基础上,结合随机森林算法快速自动提取梯田,利用CSLE模型对山区耕地土壤侵蚀进行精细定量评估,并以空间自相关分析揭示土壤侵蚀空间分布特征。结果表明:1)文山州耕地空间异质性显著,以广南、西畴、马关一线为界,呈自西向东递减格局;坡耕地占耕地总面积的79.27%,3929.57 km^(2)耕地采取水土保持工程措施,占49.79%。2)全区耕地平均侵蚀模数为5180.72 t/(km^(2)·a),其中69.99%的耕地遭受轻度及以上侵蚀;耕地总侵蚀量达4081.50万t,其中未采取保护措施耕地的侵蚀贡献率为79.40%,土壤侵蚀对坡度和高程的响应存在差异,具有明显垂直地带性特征,1000~1800m高程带侵蚀情况严峻,3073.65万t侵蚀量源于20°以上陡坡,占侵蚀总量的75.30%。3)侵蚀空间聚集特征鲜明,文山州南部和东部为侵蚀高值中心,低值中心见于中西部和北部。利用区域侵蚀因子自动提取产品和空间自相关分析能够快速完成侵蚀精细定量评估,可为区域水土保持资源精准分配和生态环境建设提供参考。

Spatial Differences of Coastal Urban Expansion in China from 1970s to 2013

As the ′economic engines′ of China,coastal cities have developed and expanded considerably since the 1970 s.In this study,information on the urban expansion process of Chinese coastal cities from the 1970 s to 2013 was retrieved.Through the analysis of stage characteristics,spatial expansion modes,influences on land use,centroid shifts,and spatial morphological evolutions,the spatial differences of coastal urban expansion were revealed.Results indicate that 1) the expansion stages of southern coastal cities are more complex than those of northern coastal cities,and the expansion ranges and speeds of the former are far higher than those of the latter;2) the spatial expansion of southern coastal cities mostly adopt a single mode,whereas northern coastal cities mostly adopt a mixed mode;3) significant spatial differences in the occupations of different land sources exist between the south and the north along the coastline;4) the centroid of coastal cities tends to migrate toward the coastline,and the centroid shift distances of southern coastal cities are clearly higher than those of northern coastal cities;and 5) the spatial morphology of southern coastal cities tends to be compact,whereas that of northern coastal cities tends to be loose.Policies have a powerful significance in terms of guiding urban expansion and its spatial differences in coastal cities.Urban expansion in coastal cities will encounter new opportunities and challenges under the guidance of new types of urbanization strategies.

Urban Expansion in China Based on Remote Sensing Technology：A Review

Urban areas and its evolution are important anthropogenic indicators and human ecological footprints, and play decisive roles in environmental change analysis, global geo-conditional monitoring, and sustainable development. China has the highest rate of urban expansion and has emerged as an urban expansion hotspot worldwide. In this paper, the progress of studies on Chinese urban expansion based on remote sensing technology are summarized and analyzed from the aspects of urban area definition, remotely sensed imagery applied in urban expansion, monitoring methods of urban expansion, and urban expansion applications. Existing issues and future directions of Chinese urban expansion are discussed and proposed. Results indicate that: 1) The fusion of multi-source remotely sensed imagery is imperative to meet the needs of urban expansion with various monitoring terms and frequencies on different scales and dimensions. 2) To guarantee the classification accuracy and efficiency and describe urban expansion and its influences on local land use simultaneously, the combination of visual interpretation and automatic classification is the tendency of future monitoring methods of urban areas. 3) Urban expansion data have become the prerequisite for recognizing the urban development process, excavating its driving forces, simulating and predicting the future development directions, and also is conducive to revealing and explaining urban ecological and environmental issues. 4) In the past decades, Chinese scholars have promoted the application of remote sensing technology in the urban expansion field, with data construction, methods and models developing from the quotation stage to improvement and innovation stage; however, an independent and consistent urban expansion data on the national scale with long-term and high-frequency(such as annual monitoring) monitoring is still lacking.

Extraction of Basic Trends of Urban Expansion in China over Past 40 Years from Satellite Images

If urban sprawl is to be avoided in China in the next ten years, it is not only crucial to understand the overall history, current status, and future trends of urban expansion there, but also these differences, and this is presently lacking. In this study, remotely sensed images with approximately 30 m spatial resolution were used to quantitatively assess the spatial and temporal patterns of urban expansion of 60 Chinese cities(1973–2013). Urban-expansion-process curves of the cities studied were drawn using annual expansion area as an indicator. Curve similarity analysis generated four basic process modes of urban expansion in China. These included cities that: 1) peaked around 2004 and then decelerated; 2) peaked around 2010 and then decelerated; 3) showed sustained acceleration, and 4) showed continued deceleration. Four basic process modes represented cities under different levels of development stage. Geographic location was found to be the most related characteristic to urban expansion process. Regional development policies at the national level in each region also showed highly temporal consistency with fluctuation characteristics of urban expansion process. Urban characteristic such as population size and administrative level were not found to be significantly related to urban expansion-process modes. Understanding the basic process-mode categories well is extremely important for future regional-balance planning and development of macroeconomic policies.

City Size Distribution and Its Spatiotemporal Evolution in China

Based on the National Land Use/Cover Database of China(NLUD-C) in the end of the 1980s(the 1980s,hereafter), 1995, 2000, 2005, and 2010, 665 cities were selected to study the size distribution and its changes of urban lands in China. In this study, the spatiotemporal evolutions of urban land size distribution as well as the influence of administrative-level on these cities were explored by combining urban spatial positions and administrative-levels. Results indicate that: 1) City size distribution using urban lands was more practical and reasonable than using non-agricultural population. 2) In the 1980s, cities with ascending urban land rank were centralized in Eastern China, specially the Changjiang(Yangtze) River Delta, the Zhujiang(Pearl) River Delta, and Beijing-Tianjin-Hebei region. Cities in Central, Western, and Northeast China mainly indicated descending urban land rank. 3) The transfer of national development focus resulted in cities with ascending urban land rank becoming evenly distributed nationwide; however, this trend was slightly centralized around Chengdu, Chongqing, and Wuhan in different periods. 4) During the 1980s to 2010, the proportion of cities with ascending urban land rank in provincial capitals, municipalities, and special administrative regions(high-level cities, hereafter) was consistently higher than those in prefecture- and county-level cities except for 2005–2010. The ranks of the majority of the prefecture- and county-level cities were mainly descending, supported by ascending; the proportion of cities with unchanged rank is small. This study breaks through the bottleneck of traditional research in the area of city size distribution by examining urban land replacing the non-agricultural population. The current study also provides scientific explanation for the healthy and reasonable development of urban land as well as the coordinated development of population urbanization and land urbanization.

土地利用强度研究进展、瓶颈问题与前景展望

本文综述了近50年来土地利用强度概念、驱动机制、方法和关键问题的发展演变,将其划分为3个研究分支:1)基于人口密度计算人类活动强度的研究;2)以传统的投入和产出强度方法为代表的研究,以及由此发展而来的复种指数、τ因子方法和技术效率方法等;3)以基于物质和能量的社会经济流通机制的土地利用强度概念模型和HANPP(human appropriation of net primary production)指标为代表的研究。针对土地利用强度研究,呈以下趋势:人口密度作为衡量土地利用强度的指标逐渐被以物质和能量社会经济流通机制为基础的土地利用强度概念模型所取代;土地利用强度评估方法也由指标构建研究倾向于解决不同自然环境下指标评估比较的关键技术与瓶颈问题,即人为与自然影响的区分;遥感数据和方法在人为活动定量识别研究中展现出巨大应用潜力和价值,可以作为土地利用强度评估方法瓶颈的突破点;农、林和草牧业等土地利用类型的强度变化机制研究极不均衡。因此,构建均衡发展的土地利用强度评估机制和指标技术体系成为未来土地利用强度研究的必然趋势。

Spatial Patterns of Irrigation Water Withdrawals in China and Implications for Water Saving

By considering numerical features, spatial variation, and spatial association, the spatial patterns of China's irrigation water withdrawals in 2001 and 2010 were explored at the regional, provincial, and prefectural scales. In addition, an overlay analysis was used to develop specific water-saving guidance for areas under different levels of water stress and with different degrees of irrigation water withdrawals. It was found that at the regional scale, irrigation water withdrawals were highest in the Middle-Lower Yangtze River region in both years, while at the provincial scale, the largest irrigation water withdrawals occurred in Xinjiang. During 2001–2010, the total of irrigation water withdrawals decreased; however, in the Northeast region, especially in Heilongjiang Province, it experienced a dramatic increase. The spatial variation was largest at the prefectural scale, with an apparent effect. The spatial association was globally negative at the provincial scale, and Xinjiang was the only significant high-low outlier. In contrast, the association displayed a significant positive relationship at the prefectural scale, and several clusters and outliers were detected. Finally, it was found that the water stress in the northern part of China worsened and water-saving irrigation techniques urgently need to be applied in the Northeast region, the Huang-Huai-Hai Plain region, and Gansu-Xinjiang region. This study verified that a multi-scale and aspect analysis of the spatial patterns of irrigation water withdrawals were essential and provided water-saving advice for different areas.

1990年—2020年粤港澳大湾区红树林动态变化遥感监测

粤港澳大湾区规划建设成为国际竞争力一流湾区和世界级城市群,资源集约利用、生态环境优美是国际湾区全面建成的重要方面之一。粤港澳大湾区拥有丰富的滨海红树林湿地,在消浪减风、保护生物多样性、净化海水及固碳等方面发挥重要生态作用。湾区红树林经历过人类开发活动的破坏,也受到了湿地保护政策指引下的良好恢复,但由于调查手段不一及监测时效性限制,目前仍缺乏标准一致的数据以科学客观地阐明粤港澳大湾区红树林湿地的历史变化与最新现状。本文利用现状监测和动态更新相结合的方法,构建了1990年—2020年期间共计4个时段标准一致、前后可比的红树林分布与变化数据库;利用高性能云计算GEE平台(Google Earth Engine),提出了一种红树林现状快速更新方法。研究表明:(1)利用GEE平台可快速提取红树林初步分类结果作为动态更新的参考底图,大大提高了目视解译动态更新方法的速度,为红树林高效管理提供及时的数据支持。(2)粤港澳大湾区过去30年的红树林面积净增长10.21 km2,红树林总体得到了较好的恢复和保护,但是在1990年—2000年期间,红树林是净减少的时期,净减少4.60 km2,主要是被新增养殖坑塘、人工用地占用;2000年以来,粤港澳大湾区红树林面积持续增加,红树林公园和自然保护区的建设对红树林存量面积起到了积极作用。(3)虽然红树林自然分布多见于潮间带,但是粤港澳大湾区2010年之后通过建设红树林公园恢复的红树林,有向内陆方向延伸建设的态势。

Urban expansion in China and its effect on cultivated land before and after initiating"Reform and Open Policy"

Urbanization in China has expanded at an unprecedented speed since the declaration of "Reform and Open Policy"and presented many challenges. Unbalanced regional development, appearance of super megacities and concomitant problems,and conflicts between urbanization and cultivated land protection are three critical problems that Chinese urbanization has to face. To develop new plans for foreseeable future urbanization in China, it is critical to understand the evolving history of cities across China. This study maps urban expansion of 60 typical Chinese cities based on large amount of remote sensing data and the labor-intensive image interpretation method, in order to understand the history of urban expansion from the 1970 s to 2013.Results showed that area of cities expanded 5.23 times compared to their area in the 1970 s. Urban expansion in China accelerated three times(1988–1996, 1999–2006, and 2009–2011) and decelerated three times(1997–1998, 2007–2008, and 2012–2013) over the 40 years. The urban area of South China expanded most significantly 9.42 times, while the urban area in Northeast China expanded only 2.37 times. The disparity among different administrative ranks of cities was even greater than(3.81 times) the differences among different regions. Super megacities have been continuously expanding at a fast rate(8.60-fold), and have not shown obvious signs of slowing down. The proportion of cultivated land among the land sources for urban expansion decreased to a small extent in the 1990 s, but cultivated land continues to be the major land source for urban expansion. Effective future urbanization needs controlling the expansion scale of large cities and reasonably developing medium and small cities, as well as balancing regional development.

中国土地利用空间格局刻画方法研究

土地利用空间格局研究是土地利用/覆盖变化(LUCC)理论和实践的基础,对土地利用空间格局进行有效刻画有利于国土资源空间优化,提升土地利用规划和管理水平。由于土地利用空间格局的研究范畴尚不明晰,目前土地利用空间格局的研究对形状、斑块分布和结构等方面关注较少,缺乏对不同土地利用类型间相互关系的研究;同时,格局指标繁多且存在较大的相关性,如何建立指标与土地利用空间格局的对应关系值得进一步研究。本文在深刻理解土地利用空间格局内涵的基础上,将面积、形状和斑块分布总结为土地利用几何特征,将结构和多样性总结为土地利用类型特征,建立了土地利用空间格局刻画指标体系,利用模糊C均值聚类分析,明确了指标与空间格局间的对应关系。结果表明,中国土地利用几何特征可以划分为简单大斑块型、复杂大斑块型、复杂小斑块型、简单小斑块型和散布型五种,不同的几何特征反映了不同土地利用类型的面积、形状和斑块分布的特点,体现了区域土地利用类型的组合关系。2010年中国土地利用共存在61种不同的结构,但主要的结构类型仅有14种,结构特征具有明显的空间聚集性,体现了不同土地利用类型的空间分异性。中国土地利用多样性以3~5类为主,占比达66.69%,其特征总体上呈现"东北、东南高,西北低"的态势。该研究丰富了土地利用空间格局研究的理论体系,填补了中国土地利用整体空间格局刻画的空白。