利用多光谱卫星遥感和深度学习方法进行青藏高原积雪判识

青藏高原积雪对全球气候变化十分重要,针对已有积雪遥感判识方法中普遍采用的可见光与红外光谱数据易受复杂地形与高海拔影响,导致青藏高原地区积雪判识精度较低的问题,提出了一种基于多光谱遥感与地理信息数据特征级融合的积雪遥感判识方法:以风云三号卫星可见光与红外多光谱遥感资料与多要素地理信息作为数据源,由地面实测雪深数据与现有积雪产品交叉筛选出样本标签,构建并训练基于层叠去噪自编码器(SDAE)的特征融合与分类网络,从而有效辨识青藏高原遥感图像中的云、积雪以及无雪地表。经地面实测雪深数据验证,该方法分类精度显著高于使用相同数据源的FY-3A/MULSS积雪产品,略高于国际主流积雪产品MOD10A1与MYD10A1,并且年均云覆盖率最低。试验结果表明该方法可有效地减少云层对积雪判识的干扰,提升分类精度。     

Snow disaster characteristics in Palongzangbu River Basin and mitigation countermeasures for road engineering

The Palongzangbu River Basin contains the highest number of maritime province glaciers in China.There are 130 glacial lakes,64 snow avalanche sites and 28 glacial debris flow gullies distributed within the basin.Snow disasters play a controlling role in the Sichuan-Tibet Highway construction,due to the terrain's special characteristics of high altitude and large height differential.Segmentation mitigation countermeasures for the Sichuan-Tibet Highway are presented based on snow disaster severity level and damage mode of the road.In the Ranwu to Midui section,snow avalanches are regional disasters, so the line should be placed in sunny slopes.In the Midui Gully to Yupu section,the line should be placed in shady slopes and at higher elevations to reduce the risk of glacial lake outburst.In the Yupu to Guxiang section,all three snow disasters are minimal.In the Guxiang to Tongmai section,glacier debris flows are the major threat,thus the road should be placed in shady slopes.     

IS THE EXTREME SNOW DEPTH RECORD FOR NORDLAND,NORWAY RELIABLE?

A snow depth of 370 cm at Dunderlandsdalen in winter 1919–20 is the largest recorded at stations in Nordland, Norway. During the period 1895–1924, the average maximum annual value there was 150.4 cm; at other stations it ranged from 38.5 to 190.1 cm. The ratio of maxima at other stations to that at Dunderlandsdalen was particularly low in 1919–20. In Nordland generally, that winter's recorded precipitation was slightly above the 30‐year average, but at Dunderlandsdalen it exceeded the average by 34%. At all stations except Dunderlandsdalen, 26 Jan.–1 Feb. was a dry period; at Dunderlandsdalen, 51.7 mm was recorded. Only one day without snowfall was recorded at Dunderlandsdalen between early January and early February, but elsewhere there were few days with snowfall. The difference in snowfall frequency and snow depth at Dunderlandsdalen in 1919–20 from values recorded elsewhere in Nordland contrasts with the relationships in other winters between 1895 and 1924. No observations were made at Dunderlandsdalen in winter 1917–18. Two of the householders there died in 1916. A change of personnel making the observations may have been responsible for the data gap and for the anomalous 1919 data. Changes made to buildings or the recording site in 1917 or 1918 may have resulted in increased snow depths as a result of drifting. Maintaining a record of climatic extremes and their environmental consequences is important. Data must be accurate. In view of this, it would be sensible to regard the validity of the 370 cm Dunderlandsdalen maximum as doubtful.     

FY-3 A/MERS I积雪制图中NDS I指标建立及积雪判识模型研究 ——以祁连山区为例

我国新型自主的极轨气象卫星风云3号A星（简称FY-3A）上搭载的中分辨率光谱成像仪（MERSI）为大面积雪监测提供了新的遥感数据源。以中国西北祁连山区为例,分析FY-3A/MERSI传感器积雪与其它地物的图谱特征差异,建立了适用于FY-3A/MERSI的归一化差分积雪指数（NDSI）,以此为基础,构建了综合利用多光谱判别指标及土地覆盖类型（LULC）定类辅助的积雪判识模型,生成250 m分辨率的日积雪制图产品。模型通过逐步逼近的树状判别结构,去除了易和积雪混淆的部分乔木林、云、云阴影、水体、湖冰、沙（盐）地等地物,并提出应考虑积雪下覆地表特性的影响,调整设定不同LULC类型的积雪判别阈值约束,实时结合区域LULC影像进行积雪的最终判定与优化。对祁连山区2010-2011年积雪季FY-3A/MERSI影像的积雪制图应用结果表明,该资料能够客观精细地反映积雪的空间分布与动态发展过程。同时利用气象台站积雪观测记录及Terra/MODIS积雪判识结果进行对比验证,结果表明基于FY-3A/MERSI建立的积雪判识模型具有较高的精度和稳定性,特别是提高了云雪区分的效能。     

2000-2014年西藏雅鲁藏布江流域积雪时空变化分析及对气候的响应研究

利用2000-2014年MODIS逐日无云积雪产品对雅鲁藏布江流域积雪特征的空间分布及变化、积雪随高程变化的规律进行了分析,并采用被动微波数据SMMR （1979-1987年）和SSM/I （1988-2008年）以及中国地面降水和气温0.5°×0.5°日值格点数据集,研究了雅鲁藏布江流域关键积雪参数对气候要素的响应等。结果表明：流域下游积雪日较大且变化剧烈;流域整体上呈显著减少的趋势;积雪日随高程的上升而增加;流域内降水呈不显著的增加趋势,而气温呈显著的增加趋势,最高气温对积雪变化影响最大;气温对积雪终日的影响明显高于积雪初日;在积雪消融期降水的增多促进了积雪的消融。     

高山区雪堆储雪的实验和模拟计算

随着全球气候变暖,滑雪运动呈现出对气候变化高度的敏感性和依赖性,各地滑雪季将不同程度地缩短。储雪作为一种应对气候变化的方法在雪务保障中逐渐得到研究和应用。在新疆阿勒泰地区吉木乃县开展储雪实验,应用谐波反应法研究了雪堆融化量与外界空气温度、太阳辐射和绝热保温结构热学性能之间的定量关系。实验中,覆盖两层绝热保温材料雪堆1的平均融化量为18.0 kg·d^-1（相当于初始质量的0.85%）,未覆盖绝热保温材料雪堆2的平均融化量为120.8 kg·d^-1（相当于初始质量的6.67%）。模拟期间观测到的雪堆1质量减少了1 438.0 kg,对应的模拟值为1 520.8 kg。谐波反应法可以为评价绝热保温结构的热学性能和提前确定储雪量提供重要的参考依据。由于许多物理过程都未考虑,导致模拟的雪堆融化量与实际融化量之间存在不确定性。雪堆所用绝热保温结构的性能可以用反射率、总传热系数、衰减度和延迟时间来衡量。     

积雪变化对中国森林凋落物分解影响研究进展

森林凋落物的分解对于维持生态系统物质循环和养分平衡具有重要意义,并受到不同积雪厚度下冻融格局的影响。冻融期（包括冻结过程期、完全冻结期、融化过程期）是冻土区凋落物分解的重要时期,该时期分解的凋落物量约占全年分解总量的一半。积雪减少通常会导致土壤温度降低、冻融循环次数增加,进而影响凋落物分解。通过综述近10年来积雪变化对我国森林凋落物分解影响的研究成果发现,积雪厚度减少在冻融期通常会抑制凋落物质量损失、碳元素释放和纤维素降解,生长季则起到促进作用,从全年来看多数表现为抑制作用。因此,冻融作用造成凋落物的物理破坏,对其分解的促进作用主要发生在后续生长季。积雪厚度减少在冻融期通常抑制氮元素释放,生长季和全年则无明显规律;磷元素和木质素目前研究还存在很大差异。最后,进一步阐述了积雪变化对凋落物分解影响研究存在的问题及未来研究发展方向。     

基于颗粒识别分析系统的碎屑流堆积物颗粒识别和统计方法研究

碎屑流堆积物颗粒识别和统计是碎屑流灾害的研究重点。文章基于图像处理孔隙（颗粒）及裂隙图像识别与分析系统（PCAS）,以贵州纳雍普洒村崩塌-碎屑流为例,结合纳雍崩塌堆积物粒径实测结果,通过阐释识别过程中阈值、孔喉封闭半径、最小孔隙面积的参数意义,研究PCAS软件在碎屑流颗粒识别与统计中的应用,并提出了颗粒识别时这些参数的选取方法。分析结果表明:（1）PCAS能自动准确地识别碎屑流堆积物颗粒与孔隙,相比人工计数更精细,所识别堆积物各区小颗粒比重较大,0～2 m颗粒粒径各区占比都在50%以上;（2）当阈值为170（像素）时能获得精细的二值图像,颗粒与孔隙得到了准确地区分;（3）不同参数取值下获得堆积物颗粒粒径分布结果不同,碎屑流堆积物颗粒识别宜采用较大的孔喉封闭半径和较小孔隙面积,当二者比值为3/30（像素）时能更好地反应颗粒粒径分布情况;（4）PCAS具有较高的可行性,统计结果显示,各粒径含量变化趋势与人工统计相近,两种统计方法各粒径占比、分布规律基本吻合,说明利用PCAS可以实现对崩塌碎屑流颗粒粒径分布的高效便捷分析。     

四川省内江市东兴区地质灾害的特征、成因及防治对策

东兴区地质灾害发育广泛,以崩塌(危岩)、滑坡和不稳定斜坡(高陡边坡)为主.其宏观发育特征有明显的分区性和季节性;灾害规模不大,但是影响人群较多,危害较大.崩塌(危岩)分布最为广泛,其主要受控因素是为水平状砂泥岩互层的差异性风化;滑坡和不稳定斜坡(高陡边坡)分布较少,降雨入渗和人工开挖坡脚是造成灾害的直接原因.在查明地质灾害成因机制和影响因素后,对不同类型地质灾害采取人工清除、填沟摸缝、砌支撑柱(墙)、避让搬迁等有针对性的措施,以达到切实可行而且经济的目的.     

Fresh snow chemistry from high mountain regions in central himalayas

Investigations of atmospheric composition in the Himalayas has been limited in both temporal and spatial scales, mainly due to difficult logistics. Ideal sites for monitoring atmospheric composition and its evolution should be free from local pollution and representative of the remote troposphere (HUEBERT et al., 1980). As the Himalayas are far removed from highly industrialized regions they provide suitable locations to monitor the chemistry of the remote troposphere and to study the evolu…