基于Google Earth Engine与机器学习的大尺度30m分辨率沙地灌木覆盖度估算

相较于降雨充沛的南方,中国北方沙地植被呈现覆盖整体偏低、空间异质性强的特点。灌木作为该区域的优势植被,对于风沙固定、食品/木材供给起着极为重要的作用。针对当前大尺度、中高分辨率干旱地区灌木覆盖度遥感产品缺失的现状,研究提出了一套通过Collect Earth样本收集器进行样本采集、利用Google Earth Engine遥感云平台的数据与计算优势开展大尺度灌木覆盖度估算的方法,并选取中国北方四大沙地之一的毛乌素沙地开展了示范应用。研究结果表明:(1)Collect Earth样本收集器可以有效地获取地面灌木覆盖度样本数据集,可以将灌木与高大乔木与草本植被进行有效区分,为灌木覆盖度估算样本集的建立打下了基础;(2)利用Landsat数据与其他辅助数据,机器学习算法可以较好地实现灌木覆盖度的估算,CART模型确定性系数R~2为0.73,均方根误差(Root Mean Square Error, RMSE)为13.66%,预测精度(Estimated Accuracy, EA)为61.8%,SVM模型R~2为0.72,RMSE为13.73%,EA为61.6%;(3)提出的基于GEE的灌木覆盖度估算体系可为我国乃至全球尺度干旱地区沙地灌木覆盖度信息提取提供有效支撑,具有较大的应用潜力。     

生物多样性野外调查地理信息管理、路线精细设计和精确导航方法

野外调查是生物多样性研究和保护的基础工作。在野外调查中, 自然地理环境十分复杂, 对相关地理信息的掌握程度、调查路线的设计质量和导航的准确性, 直接影响调查工作的成效, 甚至调查人员的人身安全。使用谷歌地球软件及相关数据转换工具, 能获取海量遥感影像和地形数据, 能将各种地理信息转换、编辑、叠加到三维地球模型上形成地理信息库, 进而精细设计野外考察路线、调查点和调查样方, 并以kml或kmz文件格式保存和分发设计结果。使用Android智能终端设备安装OruxMaps软件, 可以将多种网络地图下载制作成离线地图, 在卫星图上精确显示当前位置、记录走过的轨迹、导入kml或kmz文件进行导航。综合使用这些工具, 可以实现良好的地理信息管理、精细的生物多样性野外调查路线设计和高精度现场导航实施, 有效提高野外调查工作的成效和安全性。     

Unlocking the potential of Google Earth as a tool in invasion science

Distribution data are central to many invasion science applications. The shortage of good information on the distribution of alien species and their spatial dynamics is largely attributable to the cost, effort and expertise required to monitor these species over large areas. Virtual globes, particularly Google Earth, are free and user-friendly software which provide high-resolution aerial imagery for the entire globe. We suggest this has enormous potential for invasion science. We provide suggestions and tools for gathering data on the distribution and abundance of invasive alien trees using visual interpretation of Google Earth imagery, and propose how these data may be used for a number of purposes, including calculating useful metrics of invasions, prioritising species or areas for management and predicting potential distributions of species. We also suggest various practical uses of Google Earth, such as providing a tool for early detection of emerging invasions, monitoring invasions over time, and to help researchers and managers identify suitable field study sites. Virtual globes such as Google Earth are not without limitations and we provide guidance on how some of these can be overcome, or when imagery from Google Earth may not be fit for invasion science purposes. Because of Google Earth’s huge popularity and ease of use, we also highlight possibilities for awareness-raising and information sharing that it provides. Finally, we provide the foundations and guidelines for a virtual global network of sentinel sites for early detection, monitoring and data gathering of invasive alien trees, which we propose should be developed as part of a “citizen science” effort. There has been limited use of virtual globes by invasion scientists and managers; it is our hope that this paper will stimulate their greater use, both within the field of invasion science and within ecology generally.     

Mapping land cover and forest density in Zagros forests of Khuzestan province in Iran: A study based on Sentinel-2, Google Earth and field data

Zagros forests in western Iran have widely been destroyed because of various reasons. This study was performed to provide the land cover and forest density maps in Zagros forests of Khuzestan province using Sentinel-2, Google Earth and field data. The forest boundary in Khuzestan province was digitized in Google Earth. Sentinel-2 satellite images were provided for the study area. One 1:25000 index sheet of Iranian Mapping Organization (IMO) was selected as pilot area in the province. Sentinel-2 image of the pilot area was classified using different supervised classification algorithms to select the best algorithm for land cover mapping in Khuzestan province. In addition, to evaluate the accuracy of Google Earth data, field sampling was performed using random plots in different land covers. Field data of forest plots were applied to investigate tree canopy cover percent (forest density), as well. Classification of Sentinel-2 image in Zagros area of Khuzestan province was done using the best algorithm and the land cover was obtained. The forest density map was also obtained using a linear regression model between tree canopy cover percent (obtained from field plots) and normalized difference vegetation index (NDVI) (obtained from NDVI map). Finally, the accuracy of land cover map was assessed by some square plots on Google Earth. Results demonstrated that support vector machine (SVM) algorithm had the highest accuracy for land cover mapping. Results also showed that Google Earth images had a good accuracy in the Zagros forests of Khuzestan province. Results demonstrated that NDVI has been a good predicator to estimate tree canopy cover in the study area. Based on results, an area of 443,091.22 ha is covered by Zagros forests in Khuzestan province. Results of accuracy assessment of the land cover map showed the good accuracy of this map in Khuzestan province (overall accuracy: 91% and kappa index: 0.83). For optimum management of Zagros forests, it is suggested that the land cover and forest density mapping will be performed using SVM algorithm, NDVI, and Sentinel-2 satellite images in Zagros forests of Khuzestan province in the certain periods.     

基于Google Earth Engine的青藏高原土地沙化模拟与动态评估

青藏高原广布的高寒草原和高寒草甸,在气候变暖以及超载放牧等人类活动的影响下面临着沙化的风险,并且高寒干旱的气候条件使得高原草地生态系统极为脆弱,发生沙化后修复难度极大,对高原土地沙化时空动态评估极为必要。但目前针对大范围进行长时间快速沙化模拟监测的方法存在不足,通过对比实验构建了基于植被状况-地表反射率-土壤湿度(NDVI-Albedo-Wet)三维数据空间的沙化遥感模拟监测模型,利用Google Earth Engine遥感大数据对模型进行了代码实现,最终分析了青藏高原土地沙化时空格局的动态变化。结果揭示了青藏高原土地沙化总体缓解但问题依然存在的现象,具体体现在：(1)沙化总面积逐年减少,平均每年减少的面积达到55.13万hm²,但沙化模拟数据显示有大量的未沙化土地向轻度沙化转移。(2)沙化程度总体上趋于逆转状态,但部分区域土地沙化问题不断出现,主要分布于“一江两河”、唐古拉山南麓、柴达木盆地周边以及昌都地区,在三江源和川西甘南地区也有零星分布。     

基于Google Earth Engine平台与复杂网络的黄河流域土地利用/覆被变化分析

黄河流域在中国社会经济发展和生态安全方面的地位十分重要,过去几十年的自然演变和人类活动对其造成了深远影响,系统科学地认识黄河流域的环境格局变化是实现黄河流域高质量发展的重要前提。基于Google Earth Engine平台和复杂网络分析方法,解译并分析了1986年至2018年黄河流域的连续年度土地利用/覆被变化(LUCC)。基于1000个独立验证点进行的评估显示解译的覆盖黄河流域33年的年度土地利用/覆被数据集的7个一级类的总体准确率为82.6%,15个二级类的总体准确率为74.7%。分析结果显示黄河流域的土地系统在整个研究期间呈现出复杂的时空变化,主要的LUCC模式包括：不变或很小的变化、伴随耕地流失的城市扩张、草地恢复、果园和梯田扩张和森林增加。基于复杂网络方法对黄河流域土地系统的分析表明,土地利用/覆被转移网络中的高覆盖草地、低植被覆盖地表和落叶常绿混交林与其他地类的转移较频繁,地类节点的中介中心性和度值较高,是黄河流域土地系统中的关键地类。另外与人类活动关系密切的果园和梯田、谷物耕地和城市及建设用地节点组成了较为活跃的网络社区结构,它们都有较高的结构多样性、接近中心性,这些地...     

Fish farms at sea: the ground truth from Google Earth

In the face of global overfishing of wild-caught seafood, ocean fish farming has augmented the supply of fresh fish to western markets and become one of the fastest growing global industries. Accurate reporting of quantities of wild-caught fish has been problematic and we questioned whether similar discrepancies in data exist in statistics for farmed fish production. In the Mediterranean Sea, ocean fish farming is prevalent and stationary cages can be seen off the coasts of 16 countries using satellite imagery available through Google Earth. Using this tool, we demonstrate here that a few trained scientists now have the capacity to ground truth farmed fish production data reported by the Mediterranean countries. With Google Earth, we could examine 91% of the Mediterranean coast and count 248 tuna cages (circular cages >40 m diameter) and 20,976 other fish cages within 10 km offshore, the majority of which were off Greece (49%) and Turkey (31%). Combining satellite imagery with assumptions about cage volume, fish density, harvest rates, and seasonal capacity, we make a conservative approximation of ocean-farmed finfish production for 16 Mediterranean countries. Our overall estimate of 225,736 t of farmed finfish (not including tuna) in the Mediterranean Sea in 2006 is only slightly more than the United Nations Food and Agriculture Organization reports. The results demonstrate the reliability of recent FAO farmed fish production statistics for the Mediterranean as well as the promise of Google Earth to collect and ground truth data.     

基于时间序列分割算法的雅鲁藏布江流域NDVI(1985-2018)变化模式研究

归一化植被指数NDVI可有效表征植被生长信息,其中,长时间序列NDVI在分析全球和局部植被变化扮演重要角色。利用Google Earth Engine提供的全系列Landsat卫星数据,应用LandTrendr时间序列分割算法,讨论了雅鲁藏布江区域植被覆盖变化特征,并对植被变化模式进行判别。研究结果表明：1985-2018年间雅鲁藏布江流域,（1）NDVI总体呈现上升趋势,仅在局部区域出现下降,自上游至下游NDVI变化强度逐渐增加,1986-1990年NDVI变化最为剧烈,1991-2000年次之,2001-2017年NDVI变化强度逐渐减弱;（2）NDVI干扰变化95%集中在0-0.42之间,平均干扰时间4.96年;NDVI恢复变化95%集中在0-0.4之间,平均恢复时间12.55年;（3）NDVI干扰模式主要以持续下降为主,但2000年前的下降速率小于2000年后的下降速率;NDVI恢复模式以持续上升为主,但2000年前的上升速率大于2000年后的上升速率。     

Richer histories for more relevant policies: 42 years of tree cover loss and gain in Southeast Sulawesi,Indonesia

Understandings of contemporary forest cover loss are critical for policy but have come at the expense of long‐term, multidirectional analyses of land cover change. This is a critical gap given (i) profound reconfigurations in land use and land control over the past several decades and (ii) evidence of widespread ‘woodland resurgence’ throughout the tropics. In this study, we argue that recent advancements within the field of land change science provide new opportunities to address this gap. In turn, we suggest that multidecadal and multidirectional analyses of land cover change can facilitate richer social analyses of land cover change and more relevant conservation policies and practice. Our argument is grounded in a case study from Southeast Sulawesi, Indonesia. Using a novel analytical platform, Google Earth Engine, and open access to high‐quality Landsat data, we map land cover change in Southeast Sulawesi, Indonesia, from 1972 to 2014. We find that tree cover loss constitutes the single largest net change over the period 1972–2014 but that gross rates of tree cover gain were three times higher than gross loss rates from 1972 to 1995 and equivalent to loss rates from 1995 to 2014. We suggest the smallholder tree crop economy likely produced both forest loss and Imperata grassland restoration in this region. This case points to the need to expand rather than collapse the baselines used to study carbon and biodiversity change in tropical regions. It also demonstrates the possible utility of applying such methods to other regions.     

Tools for Address Georeferencing – Limitations and Opportunities Every Public Health Professional Should Be Aware Of

Various address georeferencing (AG) tools are currently available. But little is known about the quality of each tool. Using data from the EPIPorto cohort we compared the most commonly used AG tools in terms of positional error (PE) and subjects'' misclassification according to census tract socioeconomic status (SES), a widely used variable in epidemiologic studies. Participants of the EPIPorto cohort (n = 2427) were georeferenced using Geographical Information Systems (GIS) and Google Earth (GE). One hundred were randomly selected and georeferenced using three additional tools: 1) cadastral maps (gold-standard); 2) Global Positioning Systems (GPS) and 3) Google Earth, single and in a batch. Mean PE and the proportion of misclassified individuals were compared. Google Earth showed lower PE than GIS, but 10% of the addresses were imprecisely positioned. Thirty-eight, 27, 16 and 14% of the participants were located in the wrong census tract by GIS, GPS, GE (batch) and GE (single), respectively (p<0.001). Misclassification according to SES was less frequent but still non-negligible −14.4, 8.1, 4.2 and 2% (p<0.001). The quality of georeferencing differed substantially between AG tools. GE seems to be the best tool, but only if prudently used. Epidemiologic studies using spatial data should start including information on the quality and accuracy of their georeferencing tools and spatial datasets.     

Google trends data need validation: Comment on Durmuşoğlu (2017)

In a recent article, Durmu?o?lu (2017) proposes the use of Google Trends data to study public understanding of environmental risks and claims such data do not require validation. Here, I show this assumption is false and that some of the terms used by the author fail a validity assessment, ultimately affecting the overall value of the study. Authors and reviewers of studies aiming to take advantage of digital data should adopt an enthusiastic but critical approach, particularly with regards to key assumptions and defining methodological aspects.     

基于GEE和BRT的1984—2019年长三角生态绿色一体化发展示范区植被覆盖度变化

长三角生态绿色一体化发展示范区是国家重大战略的区域,生态环境保护是该区域发展的优先本底。探究植被覆盖度(FVC)时空变化有利于更准确地认识生态环境质量状况,对区域绿色可持续发展具有重要意义。本研究基于Google Earth Engine(GEE)云平台,分析了一体化示范区1984—2019年植被覆盖的时空变化特征和变化趋势,并基于增强回归树(BRT)量化不同生态因子对FVC的影响。结果表明: 1984—2019年间,研究区域植被覆盖整体呈先减少后增加的趋势。研究区域植被覆盖空间分布具有阶段性和区域性的特征。植被覆盖以退化为主, 相较于1984年,2019年覆被退化和改善的区域面积占比分别为49.8%和12.8%。其中,覆被退化主要发生于吴江北部、嘉善南部和青浦东北部。人类活动完全削弱了自然因素对FVC的影响。基于研究结果,提出GEE平台是实现植被覆盖常态化监测的有效工具。     

中蒙俄经济走廊荒漠化时空格局变化及其驱动因子

针对如何定量揭示中蒙俄经济走廊荒漠化的时空变化规律及其驱动因子的影响作用,在构建荒漠化程度判别标准和指标体系的基础上,基于Google Earth Engine云计算平台,在对分类回归树(CART)、支持向量机(SVM)、随机森林(RF)和Albedo-NDVI 4个模型精度进行对比分析的基础上,选取精度最高的CART模型实现中蒙俄经济走廊区域2000—2015年的荒漠化的时空变化分析。另外,在拓展线性趋势法的基础上,构建荒漠化驱动因子定量评估模型,实现气候变化和人类活动对荒漠化驱动效应的定量测度。结果表明:1)CART模型在中蒙俄经济走廊区域荒漠化信息提取精度最高,总体精度和Kappa系数分别85%和0.754;2)极重度荒漠化区域主要分布在中国内蒙古自治区西部,蒙古国的南戈壁盟、东戈壁盟和中戈壁盟;3)2000—2015年间,中蒙俄经济走廊荒漠化总体呈扩张趋势,但极重度、重度和中度荒漠化面积呈缩减趋势。其中,中国的荒漠化面积呈明显缩减趋势,荒漠化程度净改善面积为15.18万km~2,而蒙古国和俄罗斯的荒漠化现象在加重。另外,中蒙俄经济走廊荒漠化恢复区域的气候变化驱动贡献率为68.8%;而荒漠化退化区域的人类活动驱动贡献率为69.68%,表明中蒙俄经济走廊荒漠化程度加重区域的驱动作用主要来自于人类活动。     

基于Google Earth Engine的长三角城市群生态环境变化与城市化特征分析

作为城市发展的最高空间组织形式,城市群在高速城市化进程中将不可避免的对地区生态环境产生胁迫影响。如何平衡生态环境和城市化发展已经成为了值得关注的问题。以长三角城市群为例,基于Google Earth Engine云平台,通过整合日间光学遥感和夜间灯光遥感数据对生态环境状况、城市化强度以及二者在发展过程中的耦合协调特征开展了长时间序列的对比分析。结果表明：1）长三角城市群在过去近20年间的遥感生态环境指数（Remote Sensing-based Ecological Index,RSEI）稳定上升,生态环境呈现出好转的态势,但地区间的差异依旧显著;2）综合灯光指数（Comprehensive Nighttime Light Index,CNLI）能够较为准确的刻画城市群的城市化水平,研究期间内长三角城市群形成了"层次分明"、"由东向西"逐步推进的多层次的城市化格局,其中高强度城市化城市集中在上海市及其周边的无锡、苏州和嘉兴,低强度城市化城市则分布在城市群西部;3）基于CNLI和RSEI指数构建的耦合协调距离模型能够有效的识别出城市群内部城市化与生态环境的耦合协调程度,并根据象限特征将城市群城市划分为良好协调类型、初级协调类型、城市化滞后型和生态环境滞后型城市。     

Mapping of sample collection data: GIS tools for the natural product researcher

Scientists engaged in the research of natural products often either conduct field collections themselves or collaborate with partners who do, such as botanists, mycologists, or SCUBA divers. The information gleaned from such collecting trips (e.g. longitude/latitude coordinates, geography, elevation, and a multitude of other field observations) have provided valuable data to the scientific community (e.g., biodiversity), even if it is tangential to the direct aims of the natural products research, which are often focused on drug discovery and/or chemical ecology. Geographic Information Systems (GIS) have been used to display, manage, and analyze geographic data, including collection sites for natural products. However, to the uninitiated, these tools are often beyond the financial and/or computational means of the natural product scientist. With new, free, and easy-to-use geospatial visualization tools, such as Google Earth, mapping and geographic imaging of sampling data are now within the reach of natural products scientists. The goals of the present study were to develop simple tools that are tailored for the natural products setting, thereby presenting a means to map such information, particularly via open source software like Google Earth.     

A Scientific Look at Spaghetti

The United States is currently experiencing its worst man-made environmental disaster, the BP Deepwater Horizon oil leak. The Gulf of Mexico oil spill is severe in its impact, but it is only one of several global oil spill disasters in history. Students can utilize the technology of Google Earth to explore the spatial and temporal distribution of oil spills. In addition to increasing content knowledge of oil spills, Google Earth assists in developing student geographic and technologic literacy. Designed for a middle school science class, this activity may also be used in English, history, math, art, and poster-making.     

δ15N of estuarine fishes as a quantitative indicator of urbanization

Nitrogen stable isotope values (δ¹⁵N) have commonly been used as a qualitative indicator of catchment urbanization in estuaries, but no quantitative relationship has to date been established between δ¹⁵N and degree of urbanization. We sampled five species of common estuarine fish (Mugil cephalus, Acanthopagrus australis, Sillago ciliata, Gerres subfasciatus and Herklotsichthys castelnaui) of different trophic levels from six estuaries in Southeast Queensland, Australia, to test if a quantitative relationship exists between fish δ¹⁵N and urbanization in the catchment. Degree of urbanization (urban %) in the catchment adjacent to the tidal estuary was measured using polygons constructed in Google Earth images, and verified using data from the detailed Queensland Land Use Mapping Project (QLUMP). Significant linear relationships exist between fish δ¹⁵N and urban % in most cases irrespective of the section of the estuary (upper, middle and lower) where the fish were caught, with no difference in slope and elevation between lines established for different sections. Inclusion of additional fish δ¹⁵N data from other Australian estuaries in a combined multiple regression model adding trophic level as a predictor generated a significant relationship predicting increase in fish δ¹⁵N values with increasing urban % and trophic level. There is therefore a generic quantitative relationship between estuarine fish δ¹⁵N values and degree of urbanization applicable to different fish species, geographic location or feeding habit. While the generality of this relationship needs to be tested further in other geographic locations, our findings potentially improve the applicability of fish δ¹⁵N values as a quantitative indicator of urban influence in estuaries.     

Known unknowns,Google Earth,plate tectonics and Mt?Bellenden Ker: some thoughts on locality data

Latitude/longitude data in locality records should be published with spatial uncertainties, datum(s) used and indications of how the data were obtained. Google Earth can be used to locate sampling sites, but the underlying georegistration of the satellite image should be checked. The little-known relabelling of a set of landmarks on Mt Bellenden Ker, a scientifically important collecting locality in tropical north Queensland, Australia, is documented as an example of the importance of checking records not accompanied by appropriately accurate latitude/longitude data.     

Further support for the alignment of cattle along magnetic field lines: reply to Hert et al.

Hert et al. (J Comp Physiol A, 2011) challenged one part of the study by Begall et al. (PNAS 105:13451–13455, 2008) claiming that they could not replicate the finding of preferential magnetic alignment of cattle recorded in aerial images of Google Earth. However, Hert and co-authors used a different statistical approach and applied the statistics on a sample partly unsuitable to examine magnetic alignment. About 50% of their data represent noise (resolution of the images is too poor to enable unambiguous measurement of the direction of body axes, pastures are on slopes, near settlements or high voltage power-lines, etc.). Moreover, the authors have selected for their analysis only ~ 40% of cattle that were present on the pastures analyzed. Here, we reanalyze all usable data and show that cattle significantly align their body axes in North–South direction on pastures analyzed by Hert and co-authors. This finding thus supports our previous study. In addition, we show by using aerial Google Earth images with good resolution, that the magnetic alignment is more pronounced in resting than in standing cattle.