AR(p)模型在预测黑荆树林分月总生长量中的应用

本文应用时间序列分析方法,建立黑荆树林分总生长量的AR(p)模型,据以预测未来林分平均直径、平均高、蓄积量的月总生长量,其预测误差分别不超过±2％、±4％和±5％有较高的预测精度,在生产上具有一定的实用价值。     

浙江天台山常绿阔叶林次生演替序列群落物种多样性

应用物种丰富度和物种多样性指数及群落均匀度等指标对浙江省天台山常绿阔叶林次生演替序列物种多样性进行研究。结果表明：天台山次生森林从针叶林到针阔混交林到常绿阔叶林的演替过程中，物种多样性呈先变大后变小的趋势。在群落的垂直结构中，次生演替序列3种群落类型的物种丰富度和物种多样性指数的大小顺序均为灌木层（包括幼苗及幼树）＞乔木层＞草本层。在3种群落类型中，木本植物的物种多样性均大于草本植物。表4参9。     

Long-term trends in the abundance of Mediterranean wetland vertebrates: From global recovery to localized declines

Biodiversity loss is unevenly distributed in space and time. Species have reached critically low population sizes in some areas, and remain abundant in others. Similarly, some species may benefit from successful conservation plans, while others still experience severe population depletions driven by negative impacts of human activities. Although several indicators have been proposed to measure the fate of biodiversity, they are generally only implemented globally so their relevance for regional assessment is still unclear. Here, we calculated the first regional trend in the Living Planet Index for the Mediterranean wetlands (Med LPI), an indicator that summarizes the fate of global biological diversity based on the temporal trends in abundance of vertebrate populations. The Med LPI was based on 1641 vertebrate populations of 311 species recorded in Mediterranean wetlands from 1970–2008, in 27 different countries. We investigated whether trends in the Med LPI differed between eastern and western Mediterranean countries, which have different socio-economic contexts. Finally, we assessed whether and how the trend in the Med LPI was robust to changes in the number and identity of species considered. We found that, at the Mediterranean scale, the Med LPI increased steeply, which could be taken at first sight, as a general recovery of wetland biodiversity in this biogeographical region. However, we found highly contrasting spatial trends within the Mediterranean region: the average trend was positive for western and negative for eastern countries. Moreover, we showed that depending on the method used to estimate the trend in Med LPI, it can be sensitive to the number and identity of the species considered. We suggest that understanding the regional discrepancies of the trend in biodiversity indicators as well as their robustness to the species represented in the index will enhance progress assessment towards global and regional conservation strategies.     

时间分数阶扩散方程的并行高效Legendre谱方法

研究了两类时间分数阶扩散方程的并行高效Legendre谱方法,分数阶导数分别代替标准的扩散方程的二阶空间导数和一阶时间导数。空间方向采用高效的Legendre谱方法,时间方向使用了基于Fourier级数展开的Laplace数值逆,并对其参数进行了优化。给出了两类时间分数阶扩散程的数值格式和数值例子,与其他方法比,该方法数值结果更优。     

Patterns of rabies occurrence in Bhutan between 1996 and 2009

This study was conducted to evaluate incidence and seasonal trends of rabies in dogs and other domestic animals in Bhutan from 1996 to 2009. Time series analysis approach was used to determine the seasonal trend and temporal association between species-specific rabies cases in animals. A total of 814 rabies cases were reported during the 14-year period, of which cattle and domestic dogs accounted for 55% (447/814) and 39% (317/814) of the cases, respectively. The remaining 6% of the cases (50/814) were reported in horses (2%), cats (2%), pigs (1%) and goats (1%). Rabies cases were reported throughout the year with more reports during spring and summer months. The annual patterns of cases were stable from 1996 to 2005, but the incidence increased during 2006 and 2008. Fifty-nine of the 205 sub-districts reported rabies in animals from 1996 to 2009 with increased incidences in the four districts in southern Bhutan, an area located close to the border towns of India. A significant (P < 0.05) positive cross-correlation was observed between the number of cases in dogs and other domestic animals at time lags (months) 1-3 with the highest correlation (r = 0.94, P < 0.05) observed at time lag 0 (same month) indicating that the peak in rabies incidences occur in the same month when both dogs and other domestic animal cases are reported. Regression analysis predicted rabies in other domestic animal when there are reports of rabies in dogs during the previous months. This study provides useful information about the epidemiology of rabies that can be used to plan a rabies control programme in Bhutan.     

基于神经网络的土壤水分预测建模研究

分析了利用神经网络进行时间序列预测的方法及其基本工作原理,并且利用神经 网络建立了土壤水分预测模型.试验结果表明:所建立的模型具有较好的预测效果;用神经网络建立土壤水分预测模型的方法是可行的.     

河南省典型淋溶土土系划分研究

在河南省选取8个典型土壤剖面,对其进行景观特征、剖面形态学特征的描述及其理化性质的分析,并确定其在土壤系统分类(Chinese Soil Taxonomy)中的归属,结果表明,供试土壤归属于壤质混合型石灰性温性普通简育干润淋溶土、壤质混合型非酸性温性斑纹简育湿润淋溶土、黏壤质混合型温性斑纹钙积干润淋溶土、黏壤质混合型温性普通钙积干润淋溶土、壤质混合型石灰性温性普通简育湿润淋溶土和壤质混合型非酸性温性普通简育湿润淋溶土等6个土族;并在此土族基础上,根据土系划分规则和河南地区淋溶土诊断特征及构型特征,尝试提出"土系对比检索逻辑图"以阐述土系参比划分的逻辑过程。通过对比,可将8个典型土壤剖面划分为7个土系:汤阴系、汲水系、神沟系、尚庄系、侯集系、枣林系和尚店系。     

土种与土系参比的初步探讨——以海南岛土壤为例

为取得我国土种与土系研究成果的信息交流与知识共享,本文简述了土种与土系的概念及其区别、土种与土系的参比原则和依据,以海南岛土壤为例进行了参比。     

基于时间序列的胶东半岛异常降水灰色预测模型

利用胶东半岛地区24个气象观测站1981~2010年的年降水资料,根据灰色系统理论及其建模原理,建立了胶东半岛异常降水的灰色预测GM(1,1)模型,并对2011~2030年的异常降水进行了预测和比对论证。结果表明,建立的模型精度较高,通过了残差与后验差检验分析,达到优秀标准;2011~2014年的预测结果与实际吻合较好,表明结论可以用于预测;短期预测认为2015和2017年在烟台可能出现降水异常,表现为偏少;2016年在青岛和烟台可能出现降水异常,表现为偏多。