A cluster-optimizing regression-based approach for precipitation spatial downscaling in mountainous terrain

Precipitation temporal and spatial variability often controls terrestrial hydrological processes and states. Common remote-sensing and modeling precipitation products have a spatial resolution that is often too coarse to reveal hydrologically important spatial variability. A statistical algorithm was developed for downscaling low-resolution spatial precipitation fields. This algorithm auto-searches precipitation spatial structures (rain-pixel clusters), and orographic effects on precipitation distribution without prior knowledge of atmospheric setting. It is composed of three components: rain-pixel clustering, multivariate regression, and random cascade. The only required input data for the downscaling algorithm are coarse-pixel precipitation map and a topographic map. The algorithm was demonstrated with 4 km × 4 km Next Generation Radar (NEXRAD) precipitation fields, and tested by downscaling NEXRAD-aggregated 16 km × 16 km precipitation fields to 4 km × 4 km pixel precipitation, which was then compared to the original NEXRAD data. The demonstration and testing were performed at both daily and hourly temporal resolutions for the northern New Mexico mountainous terrain and the central Texas Hill Country. The algorithm downscaled daily precipitation fields are in good agreement with the original 4 km × 4 km NEXRAD precipitation, as measured by precipitation spatial structures and the statistics between the downscaling and the original NEXRAD precipitation maps. For three daily precipitation events, downscaled precipitation map reproduces precipitation variance of the disaggregation field, and with Pearson correlation coefficients between the downscaled map and the NEXRAD map of 0.65, 0.71, and 0.80. The algorithm does not perform as well on downscaling hourly precipitation fields at the examined scale range (from 16 km to 4 km), which underestimates precipitation variance of the disaggregation field. For a scale range from 4 km to 1 km, the algorithm has potential to perform well at both daily and hourly precipitation fields, indicated from good regression performance.     

汶川8.0级地震震害遥感定量化初步研究以都江堰城区破坏为例

随着遥感技术的飞速发展, 遥感在地震应急救援、 灾害调查和损失评估中的作用越来越显著。 然而由于缺乏系统的遥感震害定量研究, 使得遥感的实用性常常受到质疑。 文中叙述了遥感震害定量研究的基本思路, 提出了遥感震害指数的概念与定量分析的基本模型, 并以2008年汶川8.0级地震造成的都江堰城区震害为例, 依据建筑物震害遥感解译结果和地面震害调查结果, 进行了都江堰城区部分街区的建筑物遥感震害指数和地面调查震害指数的统计分析, 并建立了两者之间的统计关系。 初步研究结果表明, 依据高分辨率航空遥感影像和卫星雷达图像建立的建筑物遥感震害指数与地面调查震害指数及房屋倒塌率存在显著的对应性。 因此, 通过遥感震害定量化研究, 将为地震震害调查、 损失评估提供有力的方法和工具。     

西宁市区土层剪切波速随深度变化特征

通过对西宁市区67个钻孔剪切波速资料整理,采用最小二乘法进行统计回归分析,得到西宁市土层剪切波速Vs随土层埋深H的经验关系式,西宁市典型土类Vs与H的经验关系。并且通过实测剪切波速与预测剪切波速的对比,验证所得经验关系式的适用性。     

Removal of Selenium by Adsorption onto Granular Activated Carbon (GAC) and Powdered Activated Carbon (PAC)

The present work involves the study of Se(IV) adsorption onto granular activated carbon (GAC) and powdered activated carbon (PAC). The adsorbents are coated with ferric chloride solution for the effective removal of selenium. The physico-chemical characterization of the adsorbents is carried out using standard methods, e. g., proximate analysis, scanning electron microscopy (SEM), fourier transform infrared spectroscopy (FTIR), thermo-gravimetric (TGA) and differential thermal analysis (DTA), etc. The FTIR spectra of the GAC and PAC indicate the presence of various types of functional groups, e. g., free and hydrogen bonded OH groups, silanol groups (Si-OH), alkenes, and CO group stretching from aldehydes and ketones on the surface of adsorbents. Batch experiments are carried out to determine the effect of various factors such as adsorbent dose (w), initial pH, contact time (t), and temperature (T) on the adsorption process. The optimum GAC and PAC dosage is found to be 10 g/L and 8 g/L, respectively, for Se(IV) removal with C₀ = 100 mg/L. The percent removal of Se(IV) increases with increasing adsorbent concentration, while removal per unit weight of adsorbent increases with decreasing adsorbent concentration. Se(IV) adsorption onto both the GAC and PAC adsorbents is high at low pH values, and decreases with increased initial pH. The results obtained are analyzed by various kinetic models. The parameters of pseudo-first order, pseudo-second order kinetics, and Weber-Morris intra particle kinetics are determined. It is seen that the sorption kinetics of Se(IV) onto GAC and PAC can be best represented by the pseudo-second order kinetic model.     

辽宁省危险废物管理战略研究

随着环保工作的深入,特别是“非典”之后,危险废物管理已引起政府高度重视。辽宁作为国家重化工业基地,危险废物产生量加上历史积累量居全国首位;由于历史原因,存在处置能力低、技术落后、管理起步晚、严重滞后等问题,严重威胁环境安全。妥善解决这些问题是全新课题。通过借鉴国内外危险废物管理方面的理论与实践经验,开展了辽宁省危险废物管理战略性研究,从减量化、资源化、再利用及最终安全处置方面分析,围绕提高危险废物利用处置能力,加快处置设施建设,提出合理规划布局及所应遵循的原则,促进危险废物利用处置应采取的政策性措施及加强管理等建议,最终实现全省危险废物在未来“十一五”期间安全处置的目标,为建设和谐辽宁提供环境安全。     

Orthogonal Variance Structures in Lake Water Quality Data and Their Use for Geo-chemical Classification of Dimictic, Glacial/Boreal Lakes

The accumulating volumes of data collected within environmental monitoring programs facilitate the use of exploratory statistical methods of data analysis as a supplement to traditional methods of characterizing lake water quality. When principal component analysis and multidimensional scaling are applied to a matrix containing approximately 24000 samples of lake water quality variables pH, alkalinity, conductivity, hardness, color, Secchi depth and total phosphorus concentration, it is found that the total matrix variance can be approximately reproduced in an orthogonal two-dimensional base with transformations of hardness and color as best principal component representatives. This base is suggested as an empirical lake classification standard where the variance structure of subset lake populations (such as single lakes) can be referenced to the water quality standard of the generic population. Since the principal axes of the base exclusively contain inorganic and organic related variables respectively, the combined inorganic/organic characteristics of the lake can be expressed with the hardness and color variables alone. With the data matrix being large enough to produce high significance levels, and with variable ranges wide enough to represent a majority of dimictic, glacial/boreal lakes, the analysis results should be valid in many lakes throughout the world.     

山西省地震局获奖成果初步分析

利用已建成的山西省地震局科技成果库 ,对 1 990年至 1 999年山西省地震局科技成果的成果类别、任务来源及获奖人员的年龄结构等进行了统计 ,结合山西省地震局的实际 ,对当前科技人员的现状及存在的问题进行了分析 ,提出了解决这些问题的建议     

战区军事地理系列专题地图制图研究

本文论述了制作战区军事地理系列专题图的必要性和可能性:以济南战区为例,以军事地理学和地图制图学为基础理论,探讨了战区军事地理系列专题地图的图幅构成与设计的基本原理和方法;并在微机上初步研制了制作该系列专题地图的软件系统。     

中国大陆地震地下流体异常特征研究

根据《中国震例》(1966~1999年)在中国大陆发生M_S≥5.0地震188次震例资料 , 将其中与流体异常相关的126个震例中的803项流体异常按异常出现、 异常转折至发震随时间变化和异常数量随震中距距离变化分布的时、 空特征进行统计、 分析。 研究结果表明, 无论是趋势性异常还是短临异常, 在不同的时间段中, 异常数量呈现出明显的阶段性分布特征; 异常数量空间分布会随震中距的变化而不同。 在统计研究中还表明, 两者在时、 空分布方面有较好的一致性。     

应用SVM方法进行沉积微相识别

作者针对目前沉积微相中的特征提取问题,提出了应用ＳＶＭ（支持向量机）方法进行沉积微相识别的方案。该方法不是象传统方法那样首先试图将原输入空间降维（即特征选择变换）,而是设法将输入空间升维,以求在高维空间中问题变得线性可分（或接近线性可分）。因为升维后只是改变了内积运算,并没有使算法复杂性随着维数的增加而增加,因此这种方法才是可行的。所以。利用该方法更能胜任实际情况。实际处理表明该方法在小样本情况下