融合浅层特征的深度卷积神经网络互花米草遥感监测方法

基于2018年10月份黄河口入海两侧的LANDSAT8 OLI影像,提取植被指数和缨帽变换分量共9维光谱特征,构建融合浅层特征的8层深度卷积神经网络(deepconvolutionalneuralnetwork,DCNN)分类模型,开展互花米草(SpartinaalternifloraLoisel)遥感监测的方法研究,并从不同的浅层特征来具体分析互花米草的监测结果。结果表明:(1)在分类方法上,DCNN模型的总体分类精度最高,达到90.33%,与支持向量机(support vector machine, SVM)、随机森林(random forest, RF)分类器相比,精度分别提高4.78%、2.7%,互花米草的生产者精度分别提高了2.56%、0.47%,说明在滨海湿地遥感影像分类中, DCNN有着更好的应用潜力;(2)融合浅层特征后, DCNN的总体分类精度和互花米草的识别精度分别提高了0.34%和3.25%,有效提高了对互花米草的监测能力。其中,融合归一化植被水分指数(NDII)浅层特征的DCNN分类方法中,互花米草的识别精度提高最多,为2.56%,比值植被指数(RVI)次之,为2.32%。研究结果可为互花米草的监测与管理提供技术与数据支撑。     

海岸圈围对雁鸭类水鸟种群及生境的影响研究

滨海湿地是雁鸭类水鸟的重要栖息地,然而高强度的海岸圈围工程对滨海湿地生态系统和鸟类栖息地造成了显著影响,因此分析和研究海岸圈围对雁鸭类水鸟种群及栖息地的影响,对水鸟保护和滨海湿地资源可持续利用具有重要的理论和实践意义。位于长江口的横沙东滩,于2003年启动了海岸圈围促淤工程,工程的实施导致湿地生境和水鸟种群都发生着快速变化。本研究分别对2013、2015、2016和2017年,横沙东滩雁鸭类水鸟种群和生境因子的时空变化进行了调查与分析,通过拟合回归分析得到两者的相关关系,进一步探讨海岸圈围后生境的快速演替对雁鸭类水鸟种群产生的影响,从而提出海岸带圈围后水鸟栖息地的保护与管理对策。结果表明：1）海岸圈围工程的实施显著改变了滨海湿地的水鸟生境及种群特征。随着工程的推进,自然潮滩的盐沼湿地呈动态变化趋势;促淤区则随着泥沙的补给,新生了大面积的沼泽湿地;成陆区逐渐向淡水生态系统演替。伴随着生境的变化,雁鸭类水鸟种群逐渐从成陆区和自然潮滩向促淤区转移。2）不同生境类型影响雁鸭类种群数量特征的生境因子不同。未实施圈围的自然潮滩影响雁鸭类水鸟数量的主要因素是海三棱藨草/藨草群落面积;正在实施圈围的促淤区的主要影响因素是挺水植物面积、植被归一化指数（NDVI）和无植被覆盖水域面积;已圈围的成陆区主要影响因素为沉水植物面积。综上,雁鸭类水鸟偏向选择具有开阔水域、充足食物供给和相对较低植被盖度的生境类型。对于海岸圈围区域,若能合理利用圈围土地,保留部分区域为湿地,并根据水鸟的生境需求,适当营造水鸟适宜的栖息地,可在一定程度上减缓圈围工程对水鸟的影响,实现滨海湿地水鸟栖息地保护和湿地资源可持续利用的动态平衡。     

秦皇岛七里海潟湖湿地非繁殖期鸟类多样性

2016年9月至2017年4月,采用样线法和可变大小样方法相结合,对七里海潟湖湿地非繁殖期鸟类进行了调查研究。调查共记录到鸟类13目27科60种,其中雀形目鸟类10科14种,占物种总数量的23. 33%;非雀形目鸟类17科46种,占物种总数量的76. 67%。湿地水鸟数量占全部鸟类总数量的95. 11%,水鸟类群主要为鸥类、雁鸭类、鸻鹬类和鹭类。鸟类居留类型主要为旅鸟和夏候鸟,分别占物种总数量的35. 00%和33. 33%。鸟类区系方面,以古北界鸟类和广布种鸟类为主,分别占物种总数量的63. 33%和26. 67%。经统计,鸟类多样性指数、均匀度指数和丰富度指数的平均值分别是1. 86、0. 39和2. 31。七里海潟湖湿地位于东亚—澳大利西亚迁徙路线上,是重要的鸟类迁徙停歇地和觅食地,鸟类尤其是水鸟资源丰富,应当受到关注和保护。     

基于高光谱的湿地植被分类研究

光谱特征的选择对于湿地植被的识别精度和效率有直接的影响。本文以萨克拉门托-圣华金三角洲为研究区,基于Hy Map航空高光谱遥感影像数据,分析湿地植被的一阶微分和二阶微分光谱特征。在上述分析的基础上基于均值置信区间的波段选择法对一阶微分、二阶微分进行波段选择,根据获取的有效特征波段构建特征集,利用C5决策树分类算法产生规则集,并对实验区的湿地植被进行了分类研究。结果表明:湿地植被的一阶微分、二阶微分能够突出不同湿地植被光谱曲线在不同波段的增速不同,利用均值置信区间的波段选择法能够对特征波段起到降维效果,根据降维后的特征波段采用C5决策树分类算法,可以实现湿地植被在物种水平上的识别,并达到较好的分类精度。     

基于GF-1 WFV影像的黄河口湿地植被盖度估测及分析

植被盖度是植被生长状况的重要定量指标,而目前开展的植被盖度遥感估测工作主要集中在陆地区域,对河口湿地植被盖度遥感估测工作比较少见。本文基于国产GF-1 WFV遥感影像开展了黄河口湿地植被盖度估算,并结合植被类型、土壤盐度和植被指数分布状况开展了植被盖度分布特征分析,得到如下结论:（1）基于GF-1 WFV卫星影像的NDVI、SRI、SAVI、MSAVI和DVI 5种植被指数,分别建立植被盖度估测模型,其中基于NDVI、SRI、MSAVI和DVI 4种植被指数建立的多变量线性回归模型估测精度最好,其决定系数（R2）最大,为0.904,均方根误差RMSE最小,为0.14;（2）植被盖度估算模型的精度与植被盖度本身有一定的关系,其中各植被盖度回归模型中,盖度大于0.8时估算精度要优于盖度小于0.6的区域,RMSE最大相差0.04;（3）以潮滩碱蓬和潮滩芦苇为主的植被覆盖区属于低植被盖度区,盖度位于0.03~0.5,盐度在1.5 g/L左右;芦苇草甸、互花米草和柽柳灌丛植被覆盖区属于高植被盖度区,盖度位于0.8~1.0,其中芦苇草甸土壤盐度小于1.2 g/L,柽柳灌丛土壤盐度在1.4~2.0 g/L,在高盖度植被区混生有中等盖度的植被,盖度在0.5~0.8,土壤盐度在1.8 g/L左右。     

近30 a来盐城沿海盐沼时空分布演变规律及影响因素分析

盐沼是全球温带及亚热带地区的主要滨海湿地类型之一,具有重要的生态服务功能。本文以盐城沿海区域为例,利用近30 a来的6期遥感卫星图像,通过简单的植被覆盖指数(NDVI)粗提取和目视解译的方法分析了该区域盐沼的时空演变,并对引起这些变化的影响因子进行了探讨。结果表明,近30 a来研究区盐沼面积呈大幅度先升后降又小幅度上升的趋势,至2015年盐沼面积减少了约18 km²;盐沼的空间位置不断向外扩张,最大扩张距离达到了5.5 km,盐沼的宽度也随着时间不断变化;盐沼植被类型发生了不同程度的演变,景观指数表明景观破碎化程度不断加剧,獐毛、碱蓬、芦苇大量减少,单一的互花米草覆盖湿地。对盐沼时空变化起到影响的自然因素和人为因素中,围垦活动的影响最为显著。     

Rangeland privatization and its impacts on the Zoige wetlands on the Eastern Tibetan Plateau

The high Zoige Basin （Ruoergai Plateau） on the eastern Tibetan Plateau is a fault depression formed during intensive uplifting of the Tibetan Plateau. The wetland is globally important in biodiversity and is composed of marshes, bogs, fens, wet meadows and shallow water interspersed with low hills and sub-alpine meadows. Most of the Zoige wetlands have long been one of the most important grazing lands in China. Recent rangeland policy has allowed grazing, and usable wetland areas have been being legally allocated to individuals or groups of households on a long-term lease basis. Privatizafion of the wetland has impacted the Zoige wetlands in aspects of hydrologic condition, landscape and biodiversity. The uneven spatial distribution of water resources onprivatelands has led to the practice of extracting ground water, which has decreased the perched water table in Zoige. Fencing off the rangelands and grazing on expanding sand dunes have affected landscapes. Variation in the water table has led to the changes in vegetation diversity, resulting in the changes in wildlife and aquatic diversities and ecosystem processes. Making use all year round of the pasture that was previously grazed only in summer has shrunk the daily activity space of wildlife, and the newly erected fences blocked the movement of wild animals looking for food in the snow to lower and open areas. To maintain the favorable conditions of the Zoige wetland ecosystem, the author suggests that, in addition to biophysical research and implementation of conservation practices, there is an immediate need to initiate an integrated management program, increase public awareness of wetland functions and provide better training for the local conservation staff.     

湿地生态系统影响气候变化的活性气体交换通量的研究进展

湿地生态系统是影响气候变化的活性气体的重要“源”或“汇”,随着人类活动的增加和气候变化的加剧,湿地生态系统活性气体对全球气候变化的影响越来越受到关注。已开展的研究主要集中于二氧化碳(CO2)、甲烷(CH4)、氧化亚氮(N2O)和二甲基硫(DMS)等活性气体。本文综述了湿地生态系统CO2、CH4、N2O和DMS的交换通量、时空变化特征以及影响因素。然而,目前对湿地生态系统活性气体的研究还不够系统和深入,对不同类型湿地的活性气体的源汇格局,碳氮硫的耦合作用机制、人类活动和环境压迫的影响等开展系统地研究,才能定量评价其对全球气候变化的贡献,并阐明变化趋势。     

The use of local indicators of spatial association to improve LiDAR-derived predictions of potential amphibian breeding ponds

We examined whether spatially explicit information improved models that use LiDAR return signal intensity to discriminate in-pond habitat from terrestrial habitat at 24 amphibian breeding ponds. The addition of Local Indicators of Spatial Association (LISA) to LiDAR return intensity data significantly improved predictive models at all ponds, reduced residual error by as much as 74%, and appeared to improve models by reducing classification errors associated with types of in-pond vegetation. We conclude that LISA statistics can help maximize the information content that can be extracted from time resolved LiDAR return data in models that predict the occurrence of small, seasonal ponds.      

Transpiration rates of Carex Meyeriana in relation to micrometerological factors in a mountain valley wetland

In order to reveal transpiration rates of wetland plants and its relationships to micro- meterological factors in a mountain valley wetland, relative humidity, air temperature, leaf temperature, soil temperature, photo flux density and transpiration rates were measured once two hours in a Carex meyeriana wetland of the Changbai Mountain valley in dry （July） and wet （September） of 2003, respectively. Results showed that the tendency of ＂decreasing after increasing＂ was obvious in wet season. However, a relatively stable trend was observed for the transpiration in dry season.. Generally, the photon flux density of Carex meyeriana was higher in wet season than that in dry season. However, the variabilities of leaf temperature, air temperature and relative humidity were similar in both seasons. Higher transpiration rates of Carex meyeriana leaves were observed in July （varied from 40 to 150 mol.m^-2.s^-1） compared to those （varied from 7 to 14 mol.m^-2.s^-1） in September. Transpiration rates were significantly correlated with air temperature （P〈0.01）, leaf temperature （P〈0.01）, and wind speed （P〈0.05）, but correlationship between relative humidity and photo flux density was not significant （P〈0.05）.