基于神经网络的黄东海春季二类水体三要素浓度反演方法

介绍了一种基于人工神经网络的二类水体海域的三要素浓度反演方法。根据2003年春季黄东海试验中获得的高质量现场数据,建立了由现场测量遥感反射率分别反演三要素浓度的神经网络模型。反演的平均相对误差分别叶绿素32．5％,黄色物质8．9％,总悬浮物24．2％。同时分析了神经网络模型在水色反演模式应用中的稳定性。     

山东南部近海秋季长蛇鲻相对资源量的分布及其影响因素

为探究长蛇鲻(Saurida elongate)的生态习性和分布规律,并为长蛇鲻资源的合理利用与养护提供科学依据,本文根据2016年秋季在山东南部近海进行的渔业资源与环境调查数据,分析了该海域长蛇鲻的分布特征,研究长蛇鲻成体、幼体的分布差异,并利用广义可加模型(GAM)研究其分布与生物因子和环境因子的关系。结果表明,长蛇鲻成体与幼体的分布存在差异,成体分布范围广,幼体主要分布在30 m等深线及以浅水域。GAM模型的结果表明,饵料生物、底层水温、水深和底层盐度是影响长蛇鲻相对资源量分布的主要因子。成体、幼体的分布与影响因子的关系差异极显著(P<0.01)。长蛇鲻成体的相对资源量随饵料生物和底层水温的增加表现为先上升后下降的趋势,而幼体呈现一致上升趋势;成体和幼体的相对资源量随水深增加均呈下降趋势;幼体相对资源量随底层盐度增加有明显上升趋势,而盐度对成体的影响不显著。本研究认为山东南部近海是长蛇鲻的重要栖息地,水温和盐度是成体和幼体分布差异的可能原因。     

黄东海表观光学量与固有光学量经验模式研究

表观光学特性和固有光学特性是海洋光学研究领域的两大主要研究内容。文中通过2003年黄东海试验和2007年908专项海洋光学调查所获取的表观光学参数中的遥感反射率Rrs(λ)、水下遥感反射率rrs(λ)、海水透明度Zs、下行辐照度漫衰减系数Kd(490)、固有光学参数中的后向散射系数bb(λ)和水体吸收系数a(λ),建立了区域性、季节性的表观光学量和固有光学量之间的经验模式,并对模式进行了误差分析和改进。     

双源CT诊断兔急性肺栓塞实验研究

目的:应用双源CT探讨新西兰兔耳缘静脉注入明胶海绵制造急性肺动脉栓塞(PE)模型的可行性。方法:新西兰兔24只随机分为实验组(n=22只,按栓塞后检查时间分为2h、1d、3d和7d组)与对照组(n=2只)。实验组22只采用经耳缘静脉快速注射明胶海绵栓子制成急性PE模型,栓塞前后均行CTPA及肺灌注检查,检查结束后全部处死。对照组2只新西兰兔经耳缘静脉注入等渗盐水后立即处死。全部行病理检查。结果:22只实验组新西兰兔制模成功20只,其中有2只分别因为栓塞过量和麻醉过量而未取得数据,模型制备成功率为90%。以肺叶为单位,在CT图像上分析100个肺叶影像表现,可见栓塞后2h组对应的肺叶纹理稀疏12叶;1d及3d组对应肺叶呈磨玻璃改变22叶;7d组对应肺叶实变3叶。DEPI均呈现低灌注。病理检查发现:2h组相应肺组织呈鲜红色,未见结构破坏及肺泡渗出;1d组对应肺组织呈水肿、淤血、出血及炎细胞浸润改变;3d组部分肺组织轻度梗死,对应肺组织呈深红色,肺泡间隔增宽,大量炎细胞浸润;7d组肺泡腔被渗出物填充,对应肺组织呈暗红色,实变坏死。实验组共发现3只新西兰兔肺动脉管腔内含明胶海绵。结论:经新西兰兔耳缘静脉注射明胶海绵制作急性肺栓塞动物模型操作简单,成本低廉,成功率较高,是肺栓塞影像学研究较容易制作的实验模型。     

Comparison of the dynamic response of a saturated soil between two models

The dynamic response of marine sediment to ocean surface waves is treated extremely by marine geotechnical and coastal engineers. In the area of conventional hydrodynamics, the assumption of im-permeable rigid seabed has been used as the boundary of water waves theories and the waves parameters de-duced from the theory have not referred to the interac-tion of water waves and seabed. In the area of soil mechanics, many engineers regard the response as slow loading according to the Zienkewiczs[1…     

Incorporating Spatial Dependence in Predictive Vegetation Models: Residual Interpolation Methods*

Predictive vegetation modeling can be used statistically to relate the distribution of vegetation across a landscape as a function of important environmental variables. Often these models are developed without considering the spatial pattern that is inherent in biogeographical data, resulting from either biotic processes or missing or misspecified environmental variables. Including spatial dependence explicitly in a predictive model can be an efficient way to improve model accuracy with the available data. In this study, model residuals were interpolated and added to model predictions, and the resulting prediction accuracies were assessed. Adding kriged residuals improved model accuracy more often than adding simulated residuals, although some alliances showed no improvement or worse accuracy when residuals were added. In general, the prediction accuracies that were not increased by adding kriged residuals were either rare in the sample or had high nonspatial model accuracy. Regression interpolation methods can be an important addition to current tools used in predictive vegetation models as they allow observations that are predicted well by environmental variables to be left alone, while adjusting over‐ and underpredicted observations based on local factors.     

1998-2012年青藏高原TRMM 3B43降水数据的校准

运用1998-2012年青藏高原的TRMM 3B43降水数据以及气象台站实测降水数据,对比分析了青藏高原地区TRMM 3B43降水数据偏差分布规律。结果表明:(1) TRMM 3B43降水数据在青藏高原地区存在明显误差,特别是降水量大的地区和月份,偏差量较大。(2)青藏高原地区TRMM 3B43降水数据偏差分布与海拔、经纬度、降水量存在密切的关系。用偏差分布规律,加法修正法结合随机森林算法对青藏高原地区TRMM 3B43降水数据进行了校准。经过校准之后,数据精度得到显著提高,有效增加了数据的可用性,多年月平均数据决定系数R²最大可达到0.9(3、10月),最小也接近于0.5(12月),效率系数E均为正值,最大可达到90(3、10月);多年季平均和多年平均降水数据中除了第一季度结果稍差外(决定系数R²为0.58),其余数据校准效果均较好。     

Top-down and bottom-up modelling to support low-carbon scenarios: climate policy implications

Bottom-up and top-down models are used to support climate policies, to identify the options required to meet GHG abatement targets and to evaluate their economic impact. Some studies have shown that the GHG mitigation options provided by economic top-down and technological bottom-up models tend to vary. One reason for this is that these models tend to use different baseline scenarios. The bottom-up TIMES_PT and the top-down computable general equilibrium GEM-E3_PT models are examined using a common baseline scenario to calibrate them, and the extend of their different mitigation options and its relevant to domestic policy making are assessed. Three low-carbon scenarios for Portugal until 2050 are generated, each with different GHG reduction targets. Both models suggest close mitigation options and locate the largest mitigation potential to energy supply. However, the models suggest different mitigation options for the end-use sectors: GEM-E3_PT focuses more on energy efficiency, while TIMES_PT relies on decrease carbon intensity due to a shift to electricity. Although a common baseline scenario cannot be ignored, the models’ inherent characteristics are the main factor for the different outcomes, thereby highlighting different mitigation options.

Policy relevance

The relevance of modelling tools used to support the design of domestic climate policies is assessed by evaluating the mitigation options suggested by a bottom-up and a top-down model. The different outcomes of each model are significant for climate policy design since each suggest different mitigation options like end-use energy efficiency and the promotion of low-carbon technologies. Policy makers should carefully select the modelling tool used to support their policies. The specific modelling structures of each model make them more appropriate to address certain policy questions than others. Using both modelling approaches for policy support can therefore bring added value and result in more robust climate policy design. Although the results are specific for Portugal, the insights provided by the analysis of both models can be extended to, and used in the climate policy decisions of, other countries.     

Tidal effects on estuarine circulation and outflow plume in the Chesapeake Bay

The response of the Chesapeake Bay to river discharge under the influence and absence of tide is simulated with a numerical model. Four numerical experiments are examined: (1) response to river discharge only; (2) response to river discharge plus an ambient coastal current along the shelf outside the bay; (3) response to river discharge and tidal forcing; and (4) response to river discharge, tidal forcing, and ambient coastal current. The general salinity distribution in the four cases is similar to observations inside the bay. Observed features, such as low salinity in the western side of the bay, are consistent in model results. Also, a typical estuarine circulation with seaward current in the upper layer and landward current in the lower layer is obtained in the four cases. The two cases without tide produce stronger subtidal currents than the cases with tide owing to greater frictional effects in the cases with tide. Differences in salinity distributions among the four cases appear mostly outside the bay in terms of the outflow plume structure. The two cases without tide produce an upstream (as in a Kelvin wave sense) or northward branch of the outflow plume, while the cases with tide produce an expected downstream or southward plume. Increased friction in the cases with tide changes the vertical structure of outflow at the entrance to the bay and induces large horizontal variations in the exchange flow. Consequently, the outflow from the bay is more influenced by the bottom than in the cases without tide. Therefore, a tendency for a bottom-advected plume appears in the cases with tide, rather than a surface-advected plume, which develops in the cases without tide. Further analysis shows that the tidal current favors a salt balance between the horizontal and vertical advection of salinity around the plume and hinders the upstream expansion of the plume outside the bay.     

时间尺度重构EEMD-GRNN改进模型预测PM2.5的研究

合理构建PM_2.5浓度预测模型是科学、准确地预测PM_2.5浓度变化的关键。传统PM_2.5预测EEMD-GRNN模型具有较好的预测精度,但是存在过于关注研究数据本身而忽略其物理意义的不足。本研究基于南京市2014-2017年PM_2.5浓度时间序列数据,分析PM_2.5浓度多尺度变化特征及其对气象因子和大气污染因子的尺度响应,基于时间尺度重构进行EEMD-GRNN模型的改进与实证研究。南京市样本数据PM_2.5浓度变化表现为明显的天际尺度和月际尺度,从重构尺度（天际、月际）构建GRNN模型更具有现实意义;同时,PM_2.5对PM₁₀、NO₂、O₃、RH、MinT等因子存在多尺度响应效应,以其作为GRNN模型中的输入变量更具有时间序列上的解释意义。改进后的EEMD-GRNN模型具有更高的PM_2.5浓度预测精度,MAE、MAPE、RMSE和R²分别为6.17、18.41%、8.32和0.95,而传统EEMD-GRNN模型的模型有效性检验结果分别为8.37、27.56%、11.56、0.91。对于高浓度天（PM_2.5浓度大于100 μg/m³）的预测,改进模型更是全面优于传统EEMD-GRNN模型,MAPE为12.02%,相较于传统模型提高了9.03%。