Recursive residuals and model diagnostics for normal and non-normal state space models

Model diagnostics for normal and non-normal state space models are based on recursive residuals which are defined from the one-step ahead predictive distribution. Routine calculation of these residuals is discussed in detail. Various diagnostic tools are suggested to check, for example, for wrong observation distributions and for autocorrelation. The paper also discusses such topics as model diagnostics for discrete time series and model discrimination via Bayes factors. The case studies cover environmental applications such as analysing a time series of the number of daily rainfall occurrences and a time series of daily sulfur dioxide emissions.     

我国南北方城市空气环境质量对比分析

依据国家环境监测部门公布的空气环境质量日报,对27个省会城市、4个直辖市城市的2004年-2006年空气质量数据进行分析。结果显示：我国城市空气污染指数（API）时空差异明显,冬、春季节API值高于夏季和秋季,北方城市的API均值高于南方城市;PM10是大气环境的主要污染物,作为首要污染物出现所占比例南、北方城市分别是66．64％和83．97％;SO2为首要污染物所占比例南、北方城市分别是7．28％和6．19％;北方城市平均Ⅰ级和Ⅱ级天数所占百分比之和为77．00％,重污染级天数所占比例达到1．02％;南方城市平均Ⅰ级和Ⅱ级天数所占百分比之和为87．24％,重污染天数所占百分比是0．04％;城市空气质量优、良天数呈增加趋势。     

Assessment of temporal stability in risk factors of crashes at horizontal curves on rural two-lane undivided highways

Introduction: Safety of horizontal curves on rural two-lane, two-way undivided roadways is not fully explored. This study investigates factors that impact injury severity of such crashes. Method: To achieve the aim of this paper, issues associated with police-reported crash data such as unobserved heterogeneity and temporal stability need to be accounted for. Hence, a mixed logit model was estimated, while heterogeneity in means and variances is investigated by considering four injury severity outcomes for drivers: severe injury, moderate injury, possible injury, and no injury. Crash data for the period between 2011 and 2016 for crashes that occurred in the state of Oregon was analyzed. Temporal stability in factors determining the injury severity was investigated by identifying three time periods through splitting crash data into 2011–2012, 2013–2014, and 2015–2016. Results: Despite some factors affecting injuries in all specified time periods, the values of the marginal effects showed relative differences. The estimation results revealed that some factors increased the risk of being involved in severe injury crashes, including head-on collisions, drunk drivers, failure to negotiate curves, older drivers, and exceeding the speed limits. Conclusions: The hypothesis that attributes of injury severity are temporally stable is rejected. For example, young drivers (30 years old and younger) and middle-aged drivers were found to be temporally instable over time. Practical applications: The findings could help transportation authorities and safety professionals to enhance the safety of horizontal curves through appropriate and effective countermeasures.     

基于ESDA-GIS的黑龙江省区域经济差异研究

文章以黑龙江省81个县(市、区)为基本空间单元,在ArcView GIS软件和GeoDA空间统计分析软件的支持下,利用2009年黑龙江省县域人均GDP数据,对黑龙江省县域经济差异进行分析,分析结果表明:黑龙江省2009年总体经济差异较大,经济重心、人口重心和几何中心均有较大偏移,总体上黑龙江省2009年经济空间差异较大;全局Moran’s I指数为0.387 6,空间自相关特性明显,经济发展水平相似的地区在空间上集中分布。在此基础上绘制了局域Moran’s I散点图和LISA集聚图,并采用空间插值获得局域I值分异图。     

基于GIS和LSA的林地质量评价与保护分区

合理划定林地利用保护分区,对于稳定森林资源,维持生态平衡,协调经济建设与生态文明建设起到积极指导和规范作用.论文以海南省屯昌县为例,在科学评价林地质量的基础上,采用局部空间自相关分析法(LSA),探索林地质量在空间上的分布特性,耦合林地自然条件,社会经济及其空间属性,以作为林地保护分区的依据.结果显示:屯昌县林地质量分值在[5.40, 8.56]之间,整体水平中等偏上,内部差异性较小,其中,高等别林地主要分布在南坤镇与枫木镇交界处以及西昌镇西南部,低等别林地主要分布在坡心镇及南吕镇;全县林地质量在空间分布上表现出较强的正相关性,存在一定的空间聚集性特征;基于林地质量局部空间自相关分析的林地保护分区方案与林地质量评价结果基本吻合,可将全县分为重点保护区,积极保护区和一般保护区,不同保护区根据自相关类型的差异可采取不同的管护措施.该研究将空间属性纳入到林地保护分区方案的制定中,拓展了林地保护分区的手段,可为合理制定林地利用规划提供科学依据和思路借鉴.     

南疆西部不同植被类型NDVI对风沙灾害的响应

不同植被类型对风沙灾害的响应研究有助于风沙灾害的遥感监测评估和防灾减灾工作。采用传统统计学和空间自相关方法,利用MODIS的每日反射率和土地类型产品,分析了风沙灾害后南疆西部不同植被类型NDVI的变化规律及其空间分布特征。结果表明：风沙灾害后,研究区不同植被类型NDVI的差异在增大,NDVI变化值在一0．3865~0．4148之间,NDVI减小的面积占整个研究区面积的54．98％,变异系数增大值在2．44％～36．75％之间;受到风沙灾害的影响,研究区植被NDVI全局Moran’SI系数从0．7982减小到0．6786,但在空间上仍存在显著的正相关性,具有显著的空间集聚特征;由NDVI差值的局部空间自相关指标集聚图以及不同植被类型NDVI差值的空间关联区域面积百分比,可以发现落叶针叶林和裸地或低植被覆盖地受风沙灾害的影响较小,郁闭灌丛和作物受风沙灾害的影响较大。上述关于风沙灾害后南疆西部不同植被类型的变化及其空间格局的研究结果,不仅证明了防护林的风沙阻挡作用,也可以为风沙灾害的防治工作提供参考依据。     

2015-2019年成渝城市群臭氧浓度时空变化特征及人口暴露风险评价

近年来O₃污染给人类健康带来了很大威胁.本文利用2015—2019年中国环境监测总站的O₃地表监测数据,通过Global Moran''s I和Getis-Ord G^*指数等方法,分析了成渝城市群O₃浓度的时空变化特征,并利用空间插值和LandScan人口格网分布数据,基于人口暴露风险模型对该地区的O₃人口暴露风险进行了评价.结果表明：①2015—2019年成渝城市群O₃浓度总超标比例为6.9%,年际变化呈先上升后下降趋势,逐月变化呈"双峰型",5月和8月达到峰值,12月最低,季节变化表现为夏季>春季>秋季>冬季,日变化呈"单峰型",8：30左右开始升高,16：00左右达到峰值;②2015—2019年成渝城市群O₃浓度呈现出由成都及周边城市为污染中心向以成都市和重庆市为首尾的"带状"污染空间格局发展的趋势,且空间自相关性较强,逐步形成以成都市和重庆市为双中心的高浓度集聚特征;③2015—2019年成渝城市群平均O₃人口暴露风险指数处于较低风险,但空间分布差异较大,人口暴露高风险地区主要集中于成都市、内江市、自贡市、德阳市、泸州市北部及重庆市主城区,低风险地区主要集中于东西片区、绵阳市北部及成渝城市群边界,中部高风险区域有向西南方向转移的趋势,重庆市南部有高风险向极高风险转化的趋势,同时,成都市和重庆市存在显著的高风险指数集聚特征.     

基于景观指数和空间自相关的吉林大安市景观格局动态研究

以松嫩平原西部盐碱化严重的典型区域大安市为研究对象,基于2000和2010年两个时期的土地利用和增强植被指数（EVI）数据,利用移动窗口法和局部空间自相关分析,研究大安市景观格局的变化。结果表明,大安市优势景观类型是农田、草地和盐碱地,2000年农田、草地和盐碱地分别占景观类型总面积的40.20%、19.09%和19.26%,2010年分别占景观类型总面积的41.69%、18.16%和19.94%。2000-2010年,农田和盐碱地面积增加,最大斑块指数增加,景观形状变得复杂,并且盐碱地的景观连通性增强,而草地的面积减少,最大斑块指数降低,景观形状变得复杂。大安市景观格局空间分布特征明显,优势度较高的景观类型分布的区域包括草地、农田和盐碱地,景观连通性强,景观破碎化程度和异质化程度低。而各种景观类型交错分布的区域景观破碎化程度较高,景观异质性较强。大安市2000和2010年EVI的局部Moran’s I分别为0.73和0.75,在空间分布上呈现出明显的空间自相关性。2000和2010年EVI呈高-高自相关的地区大多为农田,这些区域的植被覆盖较好,EVI呈低-低自相关的地区大多为盐碱地,植被覆盖较差。大安市不同景观类型的Moran’s I和斑块密度（PD）以及斑块形状指数（LSI）呈负相关关系,和最大斑块指数（LPI）以及蔓延度指数（CONTAG）呈正相关关系。移动窗口法和空间自相关法的结合分析,有助于了解大安市景观格局的空间变化特征及植被覆盖的空间聚集规律,从而为该地区生态环境保护提供依据。     

黑龙江省粮食生产的时空格局及动因分析

采用空间自相关、聚类分析和多元回归模型等方法,论文探讨了2000 年以来黑龙江省粮食生产的时空动态及其主要影响因素.结果表明:① 黑龙江省粮食生产规模持续扩大,其中,大豆播种面积及其占农作物总播种面积的比重均经历了“上升-下降”的波动过程,玉米和水稻播种面积不断扩大,但其占农作物播种面积的比重分别在2009 和2011 年超过大豆;② 黑龙江省粮食生产具有一定的空间相关性,大豆空间集聚最显著,玉米次之,水稻最不明显;③ 粮食生产区域化和专门化态势增强,松嫩平原北部及大小兴安岭地区、松嫩平原东南部和三江平原北部分别形成了以大豆、玉米和水稻为主的粮食生产类型区;④ 多元回归分析表明,农业投入、经济环境、技术水平、市场因素是影响黑龙江省粮食生产及其结构时空变化的主要因素.     

Brood temperature, task division and colony survival in honeybees: A model

One of the mechanisms by which honeybees regulate division of labour among their colony members is age polyethism. Here the younger bees perform in-hive tasks such as heating and the older ones carry out tasks outside the hive such as foraging. Recently it has been shown that the higher developmental temperatures of the brood, which occur in the centre of the brood nest, reduce the age at which individuals start to forage once they are adult. It is unknown whether this effect has an impact on the survival of the colony. The aim of this paper is to study the consequences of the temperature gradient on the colony survival in a model on the basis of empirical data.We created a deterministic simulation of a honeybee colony (Apis mellifera) which we tuned to our empirical data. In the model in-hive bees regulate the temperature of the brood nest by their heating activities. These temperatures determine the age of first foraging in the newly emerging bees and thus the number of in-hive bees present in the colony. The results of the model show that variation in the onset of foraging due to the different developmental temperatures has little impact on the population dynamics and on the absolute number of bees heating the nest unless we increase this effect by several times to unrealistic values, where individuals start foraging up to 10 days earlier or later. Rather than on variation in the onset of foraging due to the temperature gradient it appears that the survival of the colony depends on a minimal number of bees available for heating at the beginning of the simulation.