A machine learning approach for predicting computational intensity and domain decomposition in parallel geoprocessing

ABSTRACT

High performance computing is required for fast geoprocessing of geospatial big data. Using spatial domains to represent computational intensity (CIT) and domain decomposition for parallelism are prominent strategies when designing parallel geoprocessing applications. Traditional domain decomposition is limited in evaluating the computational intensity, which often results in load imbalance and poor parallel performance. From the data science perspective, machine learning from Artificial Intelligence (AI) shows promise for better CIT evaluation. This paper proposes a machine learning approach for predicting computational intensity, followed by an optimized domain decomposition, which divides the spatial domain into balanced subdivisions based on the predicted CIT to achieve better parallel performance. The approach provides a reference framework on how various machine learning methods including feature selection and model training can be used in predicting computational intensity and optimizing parallel geoprocessing against different cases. Some comparative experiments between the approach and traditional methods were performed using the two cases, DEM generation from point clouds and spatial intersection on vector data. The results not only demonstrate the advantage of the approach, but also provide hints on how traditional GIS computation can be improved by the AI machine learning.     

Simulating methane emission from a Chinese rice field as influenced by fertilizer and water level

Natural and agricultural wetlands are considered to be the major sources of global atmospheric methane (CH₄). A one‐dimensional model was developed to simulate methane emission and used to examine the influence of various physical processes on the rate of methane emission. Three processes involved in the methane emission are implemented in the model: production, reoxidation and transport. Three transport pathways were considered: diffusion across water–air or soil–air interfaces, ebullition and diffusion through plants. These pathways are influenced by soil properties, plant growth, water‐table conditions, temperature and external inputs (e.g. fertilizer). The model was used to examine the seasonal variation of the methane emission at a rice field in Hunan, China, which was observed during a field experiment for consecutive (early and late) rice seasons in 1992. The observed seasonal variations of methane emission, and role of plants in transporting methane to the atmosphere, are captured by the model simulation. Further model applications were conducted to simulate effects of fertilizer and water‐level condition on the methane emission. The results indicate that unfermented organic fertilizer produces a higher methane emission rate than mineral fertilizer. The simulations with treatments of a deep‐water covering and constant moisture reduced the methane emission. The rice field study provides a framework for further development of the model towards simulations based on spatially distributed variables (e.g. water table, soil temperature and vegetation) at a regional scale. Copyright © 2003 John Wiley & Sons, Ltd.     

Glacier change and glacier runoff variation in the Tuotuo River basin,the source region of Yangtze River in western China

Glaciers in the Tuotuo River basin, western China, have been monitored in recent decades by applying topographical maps and high-resolution satellite images. Results indicate that most of glaciers in the Tuotuo River basin have retreated in the period from 1968/1971 to 2001/2002, and their shrinkage area is 3.2% of the total area in the late 1960s. To assess the influence of glacier runoff on river runoff, a modified degree–day model including potential clear-sky direct solar radiation has been applied to the glaciated regions of the river basin over the period 1961–2004. It was found that glacier runoff has increased in the last 44 years, especially in the 1990s when a two-thirds increase in river runoff was derived from the increase in glacier runoff caused by loss of ice mass in the entire Tuotuo River basin.     

缢蛏Sinonovacula constricta卵子发生的超微结构分析

本文描述了缢蛏卵子发生过程细胞学和超微结构的特征。在卵黄发生前期,卵母细胞胞质中存在着核质外排小体和大量的核糖体。在卵黄发生期,卵质中可见发达的内质网、环孔瓣状体和大量的线粒体及高尔基体等;卵核上有核膜泡和核膜小管的结构,并可观察到核孔数目大量增加。文中讨论了上述几种结构在卵子发生过程中的作用以及它们之间的关系。     

低空大比例尺地形图航测生产关键技术

针对传统航空摄影测量生产工艺在生产过程中遇到的影响大比例尺测图精度的问题,该文探讨了低空航测生产大比例尺地形图的关键技术。通过选用宽角相机、低航高飞行、强化影像匹配、野外布设标志控制点、优选平差模型、精化测图操作等改进方式,以提高大比例尺航测精度;组建无人飞艇低空航测系统,对提出的技术方法予以实现;最后给出了山西境内近30个县市建成区1∶500测绘生产实践的部分验证成果。     

广州地区稻田甲烷排放及中国稻田甲烷排放的空间变化

１９９３年在广州地区采用中国科学院大气物理研究所研制的自动采集和分析系统测量了稻田甲烷的排放率，首次获得了占我国２０－２５％左右水稻收获面积的华南地区稻田甲烷排放特征值。从而宏观地使我国五个主要水稻生态区的甲烷排放率都有了实测资料。稻田甲烷排放率的季节变化主要与气温及灌溉水状态的变化的较大关系，日变化规律以下午出现极大为主。本实验田的甲烷排放率低。     

新疆及邻区体波震级初步分析研究

依据《地震震级的规定（GB 17740—2017）》,分析了2009—2017年新疆地震台网所记录的新疆及邻区476次中深源地震事件,测定了13601个m_b（短周期体波震级）和12035个m_B（BB）（宽频带体波震级）的数据样本,回归分析m_b和m_B（BB）得到回归方程及量规函数,结果显示m_b和m_B（BB）相关系数为0.966,表明两者显著相关。因此,建议对于中强型中深源地震可以直接从原始速度型宽频带数字地震记录上测定长周期体波震级m_B（BB）,提高地震速报测定的速度和精度。通过震级偏差统计和台站场地响应计算,分析新疆地震台网中的XKR、HTA、ATS和KSZ等地震台站震级偏差较大的原因为砂岩、灰岩、砂土层等类型的台基放大了场地响应,说明台基类型对体波震级偏差的影响较大。与NEIC测定的体波震级对比时,发现新疆地震台网测定体波震级平均偏大0.42级,且偏差随着震源深度的增加有增大的趋势。     

Assessment of global meteorological,hydrological and agricultural drought under future warming based on CMIP6

在全球变暖背景下,分析和预测干旱的变化趋势和传播规律对于区域生态环境安全和灾害管理具有重要意义.本文基于第六次国际耦合模式比较计划(CMIP6),分析了SSP2-4.5和SSP5-8.5两种变暖情景下的气象(标准化降水指数SPI和标准化降水蒸发指数SPED,水文(标准化径流指数SRI)和农业(标准化土壤水分指数SSI)...     

GeoCube： 面向大规模分析的多源对地观测时空立方体

随着对地立体观测体系的建立,遥感大数据不断累积。传统基于文件、景/幅式的影像组织方式,时空基准不够统一,集中式存储不利于大规模并行分析。对地观测大数据分析仍缺乏一套统一的数据模型与基础设施理论。近年来,数据立方体的研究为对地观测领域大数据分析基础设施提供了前景。基于统一的分析就绪型多维数据模型和集成对地观测数据分析功能,可构建一个基于数据立方的对地观测大数据分析基础设施。因此,本文提出了一个面向大规模分析的多源对地观测时空立方体,相较于现有的数据立方体方法,强调多源数据的统一组织、基于云计算的立方体处理模式以及基于人工智能优化的立方体计算。研究有助于构建时空大数据分析的新框架,同时建立与商业智能领域的数据立方体关联,为时空大数据建立统一的时空组织模型,支持大范围、长时序的快速大规模对地观测数据分析。本文在性能上与开源数据立方做了对比,结果证明提出的多源对地观测时空立方体在处理性能上具有明显优势。