基于群智感知的街景变化检测方法

移动群智感知数据包含的图像和时空情境信息可用于检测街景图像变化,但是群智感知数据通常是低质和不规范的。为了准确检测街景发生的变化,主要解决由拍摄视角差异引起的数据低质问题。首先,针对大视差问题采用图像配准方法初步对齐图像并提取出配准特征点;然后,基于配准特征点分布从图像中提取感兴趣区域;随后,针对差值图像的误检内容,提出基于面积和多特征点的筛选法去除误检区域;最后,结合边缘检测和超像素分割算法提取完整的变化对象。与MDFNet方法进行比较,实验结果显示：当街景发生变化时,该方法的F1-measure值为55.8%,增长6%,错误率为10.8%,降低24%;当街景无变化时,该方法的错误率为2.8%,下降28%。     

Snowpack response in the Assiniboine-Red River basin associated with projected global warming of 1.0 °C to 3.0 °C

This study assesses snow response in the Assiniboine-Red River basin, located in the Lake Winnipeg watershed, due to anthropogenic climate change. We use a process-based distributed snow model driven by an ensemble of eight statistically downscaled global climate models (GCMs) to project future changes under policy-relevant global mean temperature (GMT) increases of 1.0 °C to 3.0 °C above the pre-industrial period. Results indicate that basin scale seasonal warmings generally exceed the GMT increases, with greater warming in winter months. The majority of GCMs project wetter winters and springs, and drier summers, while autumn could become either drier or wetter. An analysis of snow water equivalent (SWE) responses under GMT changes reveal higher correlations of snow cover duration (SCD), snowmelt rate, maximum SWE (SWE_max) and timing of SWE_max with winter and spring temperatures compared to precipitation, implying that these variables are predominantly temperature controlled. Consequently, under the GMT increases from 1.0 °C to 3.0 °C, the basin will experience successively shorter SCD, slower snowmelt, smaller monthly SWE and SWE_max, earlier SWE_max, and a transition from snow-dominated to rain-snow hybrid regime. Further, while the winter precipitation increases for some GCMs compensate the temperature-driven changes in SWE, the increases for most GCMs occur as rainfall, thus limiting the positive contribution to snow storage. Overall, this study provides a detailed diagnosis of the snow regime changes under the policy-relevant GMT changes, and a basis for further investigations on water quantity and quality changes.     

拉萨河流域径流对土地利用和气候变化的响应分析

拉萨河流域处青藏高原中南部,因其独特的地理位置是对气候变化较为敏感的区域之一,同时也是青藏高原人口和耕地较为密集区域。在建立SWAT模型对拉萨河流域水循环过程进行模拟的基础上,通过设置不同气候情景与土地利用状况,分析近30 a来拉萨河流域径流变化的成因,并研究径流对气候因子变化的敏感性。结果表明:①气候变化与土地利用对径流影响占比分别约为82.95%和17.05%,主要原因在于近30 a拉萨河流域土地利用情况变化不大,而气温、降水则呈显著增加趋势;②降水每增加10%,流域径流约增加11.8%,且径流对降水变化敏感性的空间差异性较小;③气温每增加1 ℃,流域总径流约增加2.5%,但径流随气温变化的空间差异性较大,其中,中上游地区径流减小0.7%,下游地区径流约增加3.6%。     

变化环境下韩江生态流量演变特征分析

选取位于粤东、闽西南地区的韩江流域为研究对象,基于CA-Markov模型对2050年流域土地利用空间格局进行预测,构建SWAT分布式水文模型,以未来土地利用情景和气候变化情景为变量进行水文模拟,分析不同情景下韩江生态流量的时空演变特征。结果表明,未来城镇化扩张将使梅江支流中上游成为韩江流域内生态流量对土地利用变化最为敏感的区域;土地利用和气候变化将导致韩江流域枯水期流量整体减小,枯水期流量对环境变化更加敏感;韩江流域生态流量变化特征将呈现从西南到东北由升到降的趋势,梅江支流中上游地区的生态流量将得到改善;梅江和汀江两大支流上游区域生态流量对气候变化更为敏感;韩江流域径流总量下降,但丰枯流量分化加剧,长期来看枯季生态流量保障风险可能进一步增大。     

阿姆河流域耕地变化及水土匹配特征分析

基于中亚阿姆河流域4个分区1990—2015年间土地利用变化和水资源等数据,运用耕地变化速率、洛伦兹曲线和基尼系数法以及水土资源匹配系数对阿姆河流域耕地变化和水土匹配状况进行了综合分析。结果表明:1990—2015年阿姆河流域耕地以146.74 km²/a的速度在扩张,且各时期、各分区扩张的速度不同;在不考虑咸海来水不断减少、生态持续恶化的前提下,绿洲区匹配状况属比较公平和相对合理状况;阿姆河流域水土资源空间分布存在明显的错位现象;全流域生态用水被农业用水严重挤占,阿姆河流域灌溉面积、灌溉用水量成为影响咸海水量及面积变化的重要因素,是一种不可持续的用水方式。     

Performance of lot-level low impact development technologies under historical and climate change scenarios

Low impact development (LID) systems have potential to make urban cities more sustainable and resilient, particularly under challenging climate conditions. To quantify performance capabilities, modeling results for an array of combinations of LIDs are described using PCSWMM at lot-level to examine performance of individual LIDs on volume and peak flow reductions. Among the four LIDs studied: rain barrel (RB), vegetative swale (VS), bioretention cell (BC), and permeable pavement (PP), PP at lot-level demonstrated the best capability for reducing surface runoff volumes and peak runoff rates under historical weather conditions, while BC showed similar capability for reduction of runoff volumes but minimal peak flow reduction. With PP as the controlling method at lot-level, the maximum percentage reduction of runoff volume for a 2-year storm is 58% whereas for a 100-year storm, the runoff volume reduction is 20%. These results mean the extent of flooding that may arise from the 100-year storm is reduced, but not eliminated. Effectively, the 100-year storm volumes with LID are devolved to have flooding equivalent to a 25-year storm. Under climate change scenarios, performance for all LIDs declined at various levels, where BC was the most resilient LID for a climate change scenario, such that projected 2-year or 5-year storms with climate change will have its impact devolved with LID in place, to result in similar volumes and peaks without LID under historical conditions. Furthermore, even with an assembly of lot-level LIDs distributed throughout the community, there is not attenuation to substantial degrees of flooding for major events, but there can be effective control for water quantity for small (2- to 5-years in particular) storm events.     

基于CMIP5的中国西北地区暖湿化演变研究

为了解中国西北地区历史与未来时期的气候演变趋势,以CMFD再分析数据集及CMIP5全球气候模式输出数据为基础,从时空维度详细分析了西北地区历史气候的演变特征并对区域未来气候发展进行了预估。结果表明：1979-2018年西北地区的气温、降水量以及近地面空气比湿均呈现出明显的增加趋势,且未来极有可能继续增加;此外,降水量在年内的分配过程越趋均匀化,且降水结构在2001年左右发生了较大突变。整体来看,西北地区的历史气候演变显示出高度的暖湿化倾向。未来情景下,西北地区的气候将继续朝暖湿化方向发展,同时降水量的年内分配不均匀性也将进一步减小,且在RCP8.5排放情景下表现的更为突出。本研究可为西北地区气候变化的影响评估提供参考,同时也为水利、农业、气象和生态等部门的决策提供科学依据。     

大凌河流域典型水生植物微生物群落吸附特征试验分析

本文采用高通测序方法对大凌河流域五刺金鱼藻、格菱、狐尾藻、水蓼及芦苇5种水生植物不同季节的微生物群落进行试验分析,分析表明:藻类密度在沉水植物表层要高于浮叶植物表层,微生物群落在浮叶植物根部的结构与其表面存在明显差异。各水生植物的优势生物群落依次为变形菌、梭菌、柔膜菌、鞭脂杆菌、杆菌、疣微菌门,且各微生物群落在季节丰度上存在明显差异性。由此可见,水环境对生物群落的影响要高于植物类型。研究成果可为大凌河流域生态修复工程提供重要理论依据。     

The effects of long-term VDT usage on the nature and incidence of asthenopic symptoms

An opportunity arose in 1985 to become involved in a transition of working practice from hard copy to VDT. Over a two-year period, 161 VDT users and 65 control subjects in the same office environment were regularly and routinely examined for symptoms of asthenopia. The analysis of data shows that there are no significant differences in type, number and frequency of the work-related symptoms between VDT users and non-VDT users. It appears that reporting of symptoms within the group may be random, although certain symptoms do appear more frequently than others. Additionally, it appears that there is a significant difference between male and female users in the frequency of symptom reporting.