四川茂县叠溪镇新磨村滑坡特征与成因机制初步研究

长期以来,中国四川省茂县地区受地质、地形条件和构造活动的影响,滑坡等地质灾害频发,给人民的生命财产和公路等基础设施安全带来了巨大的威胁,因此需要对滑坡隐患区域进行有效识别和监测。以时序哨兵1号A、B卫星(Sentinel-1A/1B)影像为数据源,利用时间序列合成孔径雷达干涉测量(interferometric synthetic aperture radar, InSAR)技术对茂县岷江河谷区段的潜在滑坡隐患开展识别监测,对重点区段进行了分析,同时分析探讨了InSAR滑坡监测中不同轨道数据的视线方向形变测量灵敏度差异。从实验结果中成功探测识别出了茂县岷江河谷沟口乡至石大关乡段的20余处滑坡隐患,现场实地考察验证了识别结果的准确性,证明了时序InSAR方法在高山河谷区滑坡隐患识别监测中的有效性。     

利用S-G滤波进行MODIS-EVI时间序列数据重构

采用S-G滤波方法对MOD13A2-EVI自2001~2007年间数据进行时间序列重构,以达到去云、消除离异值的目的,从而提高数据质量及可信度。对比重构前后数据发现,重构后的EVI数据在空间上更加一致,在时间维度上时序数据年间变化更加稳定。     

B) Organisations internationales

Abstract

A modification of Svanidze's fragments method was applied to generate synthetic samples of monthly inflows to the La Angostura and Malpaso dams, located on the Grijalva River, in the state of Chiapas, Mexico. This modification allows generation of the monthly runoff of several dams in a series and preservation of their cross-correlations. Comparing statistics of generated and historical series, a good agreement between them was observed, especially for the autocorrelations between consecutive months and for cross-correlations between both dams, except for slight differences in the standard deviation and the skewness during the dry period.     

基于卡尔曼滤波重建的GPS点位修正

由于单点定位的结果受卫星星历误差、卫星钟误差以及卫星信号传播过程中大气延迟误差的影响较为显著,因此解算出的定位结果在真值附近上下浮动。文中采用ARMA模型建立卡尔曼滤波的观测方程和状态方程,并对定位结果进行滤波;采用一次滤波后的坐标值作为初值,建立ARMA模型并二次滤波。实验表明,滤波有效防止了定位结果偏差过大情况的发生,使滤波收敛值与准确值最大偏差不超过3cm,表明采用一次滤波后的坐标值建立的模型更为合理,从而为单点定位结果的时间序列模型的建立提供一种新方法。     

基于Sentinel-2时间序列数据及物候特征的棉花种植区提取

棉花是中国重要的经济作物,在新疆大面积种植。及时、准确获取棉花种植面积,对农业政策制定与农业经济发展有重要意义。以渭干河—库车河三角洲绿洲棉花为主要研究对象,利用2018—2020年（1景/1月）36景哨兵2号（Sentinel-2）数据,构建归一化植被指数（Normalize difference vegetation index,NDVI）和红边归一化植被指数（Red edge normalize difference vegetation index,RENDVI₇₈₃）时序数据;采用Savitzky-Golay（S-G）滤波法对时序数据进行平滑、重构并提取11个物候特征;利用袋外误差法对11个物候特征进行特征优选;在此基础上利用重构后的时序数据（NDVI Fit）、（RENDVI₇₈₃ Fit）、物候特征（RENDVI₇₈₃ Ph）、物候特征优选组合构建6种不同的特征数据集,利用随机森林分类（RFC）方法分别进行分类和提取,并采用最大似然分类方法和支持向量机分类方法对分类效果进行验证。结果表明：（1） NDVI和RENDVI₇₈₃时序数据变化趋势较为一致,棉花在5月（苗期）到8月初（开花盛期）有明显的上升趋势,在8月末至9月（花铃期）达到峰值。相比NDVI,红边波段构成的RENDVI₇₈₃时序曲线峰值从0.7提高到0.9,棉花区分效果更佳。（2） 11个物候特征中拟合函数最大值、生长季长度、生长季振幅、生长季结束、生长季大积分和生长季小积分对分类的贡献性最大,重要性得分分别为1.43、1.40、1.23、1.16、1.02和1.01。（3） RFC方法对特征数据集（RENDVI₇₈₃ Fit+物候特征优选组合）分类精度最佳。总体精度和Kappa系数分别为92.20%和0.92。（4） 研究区内棉花分类精度达到了91.02%,种植面积约为3424 km²,占研究区总面积的24.67%。     

Assessing Trends in Nutrient Concentrations in Coastal Shelf Seas: a Case Study in the Irish Sea

Single point time-series data collected in the Irish Sea since 1954 by Allen et al. (1998) have been re-evaluated using more recent data and areal winter surveys of the Celtic and Irish Seas. Survey data illustrate the heterogeneity of nutrient distributions with N (≈28 μM) and P (2·0 μM) enrichment of the eastern Irish Sea relative to the western Irish Sea (7–8 μM N and 0·7 μM P). Salinity nutrient relationships demonstrate distinct regional differences throughout the two seas. Trends in the Isle of Man time-series are unlikely to have resulted from changes in analytical procedures and personnel. Concerns over data quality cannot be resolved and there are no independent data to validate the time-series. Including more recent data shows P has declined since the late 1980s and N concentrations have remained stable since the mid 1970s. The absence of a Si trend is consistent with limited anthropogenic influence on riverine Si concentrations. Trends in N and P are consistent with changes in riverine concentrations of these two nutrients and the biogeochemical processes controlling their cycling in shelf waters. Denitrification is the most likely reason for offshore Irish Sea concentrations of N being lower than expected.     

A TIME SERIES ANALYSIS OF STATE GASOLINE DEMAND, 1975–1980 *

ARIMA time series model building techniques are used to construct fifty-one state gasoline demand models based on monthly data for the period of January, 1975 to July, 1960. Statistically satisfactory models are obtained for all states. Price elasticity estimates are >0 for all states. All but four are statistically significant at the 0.05 level. The significant price elasticity estimates range from ?0.138 to ?0.377, with most clustering about ?0.2. Estimates of state gasoline supply shortages for May, June, and July, 1979 are also presented which range from 0 to 8 percent of normal consumption for the three-month-period.     

Stream water geochemistry from selected catchments on the Kola Peninsula (NW Russia) and in neighbouring areas of Finland and Norway: 2. Time-series

Stream water composition, measured weekly for 8–9 months in 1994 in three arctic catchments on and around the Kola Peninsula (Russia, Finland and Norway), is presented in the form of time-series. In all three catchments, snowmelt causes a major dilution of the stream water, as reflected by marked dips in electrical conductance. In the most polluted catchment (C2), the snowmelt flood (the major hydrological event at these latitudes) is reflected in the stream water by a pH dip and a pulse in technogenic heavy metals (Cu, Ni, etc.), Al and S. This results from melting of the snow laden with heavy metals and sulphate, and from leaching of the topsoil layer. In the most pristine catchment (C8), snowmelt causes no heavy metal pulse (remote location) but yields an increase in stream water Al (acidic lithology/overburden). In the intermediate catchment (C5), very subdued heavy metal and S increases are noticeable in the stream water, whilst its pH increases steadily until summer (basic lithology). Some elements (Cl, S) may be mobilised out of the snowpack before its complete thawing and reach the stream 1–2 weeks ahead of the heavy metals. The substrate (soil, overburden and bedrock) of a catchment controls to a large extent its ability to buffer acid inputs.     

基于客源地的聚类-ARIMA模型的短期旅游需求预测--以天津欢乐谷主题公园为例

大多数旅游需求预测研究是基于目的地游客总数或消费总量开展的,尚未按不同的旅游目的或客源地细分进行预测.以天津欢乐谷主题公园为案例地,选择2014年第40周到2015年第26周为研究时段,利用通信大数据,提出了一种面向客源地的聚类-ARIMA组合预测模型.通过对不同客源地的时序数据进行聚类,选取各类别中的代表性客源地分别构建ARIMA预测模型.结果表明:对欢乐谷主题公园各客源地分别建模与聚类后通过6个代表客源地建模得到的结果一致;后者可以降低80%的预测成本.该方法具有较高的预测精度和较低的计算成本,适合面向客源地的短期旅游需求预测,可为旅游目的地提供更具针对性的旅游需求管理、分析与决策支撑.