青藏高原河谷地区历史时期耕地格局重建方法探讨——以河湟谷地为例

青藏高原受其特殊自然地理环境条件的限制,耕地主要分布在自然环境条件相对优越的河谷地区,人为因素对耕地分布范围的作用和影响极其微弱,尤其是在历史时期生产力水平较低的前提下,耕地的空间分布主要取决于土地的宜垦程度。本文将影响青藏高原河谷地区耕地分布的因子按其性质分为限制性因子和非限制性因子,并以此为基础排除了高原河谷地区不适宜耕作的地区,在适宜耕作的地区根据土地的宜垦程度,按"先优后劣"的原则将历史时期的耕地数据分配到空间上。选取青藏高原农业发展历史悠久的河谷地区之一河湟谷地作为实例,重建该区1726年耕地空间格局。将重建结果与已有的M模型重建结果进行对比分析,两者重建的耕地在空间分布上呈现出一致性,但重建结果在垦殖范围与垦殖强度上存在一定的差异;M模型的重建主要是以现代耕地分布格局为基础重建,忽略了现代耕地空间分布受现代农业技术的影响;而本文模型则是从低生产力水平前提下影响历史时期耕地分布的因子出发,重建结果更具合理性。     

基于SLM模型的中国碳排放格网化空间分布模拟

以美国橡树岭国家实验室二氧化碳信息分析中心的中国碳排放数据、统计年鉴中的分省化石能源消费与水泥产量数据为基础,选取NDVI植被指数、DMSP/OLS夜间灯光数据、GDP、工业总产值、人口5个指标,对中国2000年及2010年的分省碳排放结果采用5km×5km格网,运用空间滞后回归模型(SLM)进行空间分布模拟,并对空间模拟结果进行误差分析及纠正,获得2000年及2010年地均碳排放强度空间分布图。结果表明,中国地均碳排放强度存在以下特征:区域差异分布特征明显,主要表现为从东部地区向中西部地区逐渐变小;存在几个明显高值热点及低值冷点区域;随着时间的推移,地域差异特征越发明显,冷点和热点区域分布特征更加显著;2010年各地区地均碳排放强度明显大于2000年。     

基于随机森林的暴雨灾害人口损失预估模型及应用

基于2009—2017年的广西县级暴雨灾情记录,综合考虑致灾因子、孕灾环境和承灾体因素选取7个解释变量,运用随机森林算法,构建暴雨灾害人口损失预估模型;并以精细化网格降水实况分析和预报产品驱动模型,预估是否发生人口损失。研究结果表明:模型训练样本及测试样本的分类准确率均在90%以上,致灾因子(降水情况)是最主要的解释变量,重要性从大到小依次是前10 d降水距平百分率、过程最大日雨量、最大小时雨量和短时强降水频次。应用智能网格降水产品对广西地区近两年的暴雨灾害过程进行回报试验,准确率超过70%。     

中央气象台决策气象服务智能移动终端的设计与实现

中央气象台决策气象服务智能移动终端(简称为“决策服务App”)是在发展智慧气象的大背景下,针对桌面应用、网站等传统天气预报业务系统存在时效性、移动性不足的问题,利用最新的移动互联技术,面向决策服务领导和一线决策服务人员设计研发的手机应用程序。在研发过程中实现了基于WebGIS和OpenGL ES技术的渲染引擎、基于大数据分析技术的用户行为统计,以及高安全性等多项关键技术。在此基础上实现了包括基于精细化智能网格的城市天气预报、决策服务产品、实况监测及其统计查询,以及预警预报在内的主要功能。决策服务App在日常天气预报业务、国家重大会议与活动,以及特大灾害的气象保障服务中提升了中央气象台和各级台站的精细化分析和决策服务水平。     

Inference of a detailed P–T path from P–T pseudosections using metapelitic rocks of variable composition from a single outcrop, Shackleton Range, Antarctica

Generally, P–T pseudosections for reduced compositional systems, such as K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O, Na₂O–K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O and MnO–K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O, are well suited for inferring detailed P–T paths, comparing mineral assemblages observed in natural rocks with those calculated. Examples are provided by P–T paths inferred for four metapelitic samples from a 1 m² wide outcrop of the Herbert Mountains in the Shackleton Range, Antarctica. The method works well if the bulk composition used is reconstituted from average mineral modes and mineral compositions (AMC) or when X‐ray fluorescence (XRF) data are corrected for Al₂O₃ and FeO. A plagioclase correction is suitable for Al₂O₃. Correction for FeO is dependent on additional microscopic observations, e.g. the kind and amount of opaque minerals. In some cases, all iron can be treated as FeO^tot, whereas in others a magnetite or hematite correction yields much better results. Comparison between calculated and observed mineral modes and mineral compositions shows that the AMC bulk composition is best suited to the interpretation of rock textures using P–T pseudosections, whereas corrected XRF data yield good results only when the investigated sample has few opaque minerals. The results indicate that metapelitic rocks from the Herbert Mountains of the Northern Shackleton Range underwent a prograde P–T evolution from about 600 °C/5.5 kbar to 660 °C/7 kbar, followed by nearly adiabatic cooling to about 600 °C at 4.5 kbar.     

基于边界替代的人口数据空间化方法研究

首先将山东省各县市按级别分类,利用回归方法得到不同级别县市的城镇居住密度与农村居住密度,然后根据乡镇驻地位置,构造Voronoi图来替代乡镇边界。在此基础上利用各乡镇人口数据与居民点数据计算得到各 Voronoi图形的居住密度,并将该密度赋予每个图形的发生点(即乡镇驻地位置),而后利用反距离插值算法,并结合窗口移动平均处理,得到全省的居住密度分布图。使用居住密度分布图和居民点分布数据,计算人口分布数据。并根据每个县市的图上汇总人口和统计人口,计算得到各县市的调整系数,使用该系数对每个栅格的人口数进行调整,保证了各县市总人口不会出现误差。最终得到山东省1km×1km栅格的人口分布数据。最后选取某些样本县, 在样本县内对比每个乡镇图上的人口与实际统计人口,结果表明该人口分布数据有较高的精度。     

基于网格计算环境下的空间信息多级格网研究

本文提出一种适应网格计算环境下的空间信息多级格网(SIMG，Spatial Information Multi-Grid)技术，通过将网格计算技术与地学格网相结合，提出SIMG的构想并探讨其关键技术的实现；通过研究格网内细部地物表示的精度，力求建立SIMG的划分标准，从而建立我国空间信息多级格网的体系结构。     

Thermodynamic modelling of the reaction muscovite+cordierite→Al2SiO5+biotite+quartz+ H2O: constraints from natural assemblages and implications for the metapelitic petrogenetic grid

The reaction muscovite+cordierite→biotite+Al₂SiO₅ +quartz+H₂O is of considerable importance in the low pressure metamorphism of pelitic rocks: (1) its operation is implied in the widespread assemblage Ms + Crd +And± Sil + Bt + Qtz, a common mineral assemblage in contact aureoles and low pressure regional terranes; (2) it is potentially an important equilibrium for pressure estimation in low pressure assemblages lacking garnet; and (3) it has been used to distinguish between clockwise and anticlockwise P–T paths in low pressure metamorphic settings. Experiments and thermodynamic databases provide conflicting constraints on the slope and position of the reaction, with most thermodynamic databases predicting a positive slope for the reaction. Evidence from mineral assemblages and microtextures from a large number of natural prograde sequences, in particular contact aureoles, is most consistent with a negative slope (andalusite and/or sillimanite occurs upgrade of, and may show evidence for replacement of, cordierite). Mineral compositional trends as a function of grade are variable but taken as a whole are more consistent with a negative slope than a positive slope. Thermodynamic modelling of reaction 1 and associated equilibria results in a low pressure metapelitic petrogenetic grid in the system K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O (KFMASH) which satisfies most of the natural and experimental constraints. Contouring of the Fe–Mg divariant interval represented by reaction 1 allows for pressure estimation in garnet‐absent andalusite+cordierite‐bearing schists and hornfelses. The revised topology of reaction 1 allows for improved analysis of P–T paths from mineral assemblage sequences and microtextures in the same rocks.     

SEARCHING METHODS FOR RAPID GRID INTERPOLATION*

A widely used method for grid interpretation from a set of discrete sample points is based on the inverse distanceweighted model. This method requires the identification of k-nearest sample points for each grid point, probably the most time-consuming portion of the grid interpolation process. Three searching methods for identifying the k-nearest sample points are discussed and compared. Two of these methods were consistenty faster without sacrificing accuracy than the brute-force searching method, and one method was from one to over 90 times faster than the brute-force searching method.     

A high resolution forecast model of storm surge inundation

In order to forecast storm surge inundation, a two-dimensional model is established. In the model, an alternating computation sequence method is used to solve the governing equations, and the dry and wet method is introduced to treat the moving boundary. This model is easy to use. It has a friendly input interface and Arcview GIS is used as the output interface. The model is applied to the Shantou area to simulate the storm surge elevations and inundations caused by Typhoons 6903 ane 0104 using the same relevant parameters. The calculated results agree well with the observations.