基于多源遥感数据的高寒草地生物量反演模型精度——以夏河县桑科草原试验区为例

基于MODIS、Landsat-8 OLI和HJ-1A/1B CCD卫星遥感资料,结合2013-2014年甘南州夏河县桑科草原试验区野外实测数据,建立了高寒草地地上生物量遥感反演模型,筛选出基于不同遥感资料植被指数的生物量最优反演模型,比较分析了生物量最优模型的空间效应。同时,分析了2000-2013年基于MODIS植被指数估算的试验区产草量的年际变化特征。结果表明,草地生物量最优反演模型为基于Landsat-8 OLI NDVI数据的对数模型(y=727.54lnx1+495.23,R2=0.772,RMSE=31.333 kg DM·hm-2);在30和250 m空间分辨率下,基于MODIS NDVI及EVI、Landsat-8 OLI NDVI和HJ-1A/1B CCD NDVI最优模型估算的生物量均高于实测生物量,其中Landsat-8 OLI NDVI数据估算的草地生物量与实测生物量值最接近;2000-2013年试验区草地总生物量整体上具有显著增加的趋势(R2=0.590 7,P0.001),平均增加速率达50.57 kg DM·hm-2·a-1。     

中国北方草地覆被的HJ星NDVI校正研究

NDVI时间序列数据已经得到了广泛应用。中国的环境与灾害监测预报小卫星(HJ星)装载的4部CCD传感器能够满足在中国及其周边地区植被动态监测时间与空间分辨率两方面的需求,基于其建立的NDVI时间序列产品对于研究该区域的植被变化状况具有重要意义。然而波段的差异对不同传感器获取的NDVI具有非常显著的影响。本研究立足于对HJ星4部CCD传感器以及LandSat卫星TM、ETM+传感器的波段分析,利用同时相遥感影像对获取并推算6部传感器NDVI在中国北方中温带干旱半干旱草地的NDVI交叉校正系数,并通过野外实测光谱数据计算等效NDVI对研究结果进行验证。结果表明,HJ星4部CCD传感器间获取的NDVI差异能够达到5%,本研究中得到的传感器间NDVI交叉校正系数能够将该差异有效减小到2%以内。同时证明了传感器中心波长对NDVI的影响大于波段宽度对NDVI的影响。     

玛曲高寒草甸放牧强度的遥感监测

为给玛曲县高寒草甸草-畜平衡和草地资源的可持续利用提供科学依据,本研究利用遥感技术,结合2016年地面实测数据和同时期的MODIS影像资料,分别建立了研究区草地地上生物量和植被指数(NDVI)、相对放牧强度间的回归模型以及不同相对放牧强度和NDVI值之间的幂回归模型。结果表明,6-9月调查点NDVI和地上生物量间存在较好的相关性,即随着NDVI值的增加生物量呈上升趋势,且9月份的拟合(R2=0.510 2)达到了极显著水平;实测地上生物量和相对放牧强度之间有很好的拟合关系(R2=0.965 1),随着放牧强度的增加地上生物量呈下降趋势;相对放牧强度和植被指数间相关性较好(R2=0.631),即随着NDVI值的变小,相对放牧强度逐渐增强,当相对放牧强度增强到一定程度时,NDVI对其响应不灵敏。     

植被光合有效辐射吸收分量及最大光能利用率算法的改进

植被冠层对入射光有效辐射的吸收系数 (FPAR) 和最大光能利用率是NPP模拟的重要参数。利用光量子仪实测了典型草原植被类型关键生育期光合有效辐射值,根据NDVI与FPAR之间的良好相关性,建立了2个FAPR估算模型。结果表明,实测的FPAR值与模型估算的FPAR之间的相关性均达到极显著水平,2个模型的估测值之间没有显著性差异。因此,本研究中采取求均值后计算了我国草原的FPAR,并与NASA-FPAR产品进行了嵌套和互补,形成全国完整的草原FPAR空间分布图。同时,基于归一化植被指数和叶面积指数,以中国草原18大类分类结果为基础单元,对中国草原18大类的最大光能利用率进行调整,调整后的最大光能利用率输入NPP模型后,改善了NPP估算精度,在大类尺度上,总体预测精度为85.71%。     

基于UAV技术和MODIS遥感数据的高寒草地盖度动态变化监测研究—以黄河源东部地区为例

利用黄河源东部地区野外实测样地数据和MODIS卫星遥感资料,结合农业多光谱相机(agricultural digital camera,ADC)、普通数码相机(Canon)、无人机(unmanned aerial vehicle,UAV)等设备获取的高寒草地盖度数据,构建了基于MODIS NDVI、EVI的草地盖度反演模型,比较分析了不同草地盖度监测方法的精度,确立了黄河源区草地盖度遥感监测的最优反演模型,并分析了研究区近16年草地植被盖度的动态变化。结果表明,1) MODIS NDVI与基于UAV相片计算的草地盖度间的相关性优于MODIS EVI,而MODIS EVI与ADC和Canon照片计算的草地盖度之间的相关性则优于MODIS NDVI;2) 就Canon和ADC方法构建的草地盖度反演模型而言,前者精度远高于后者,普通数码相机方法更适宜于高寒草地植被盖度的估算;3) 对比分析两种植被指数与Canon相机、ADC和大疆(DJI)无人机航拍(航高30和100 m两种方法)相片计算的草地盖度之间的关系表明,MODIS NDVI对航高30 m UAV航拍相片计算的盖度数据的响应最敏感,基于UAV航高30 m的相片和NDVI构建的草地盖度反演模型最优;4) 黄河源东部地区2000-2015年间草地盖度稳定不变的区域达71.46%,多分布在东南部;呈增加趋势的区域占研究区草地面积的22.01%,由西向东、由北向南增加幅度呈减少趋势;盖度减少区域零星分布在黄河源北部和南部的部分地区,仅占研究区草地面积的6.53%。     

基于Landsat8 OLI遥感影像的延河流域土壤水分反演研究

在生态环境脆弱、水土流失严重的延河流域,土壤水分的遥感反演对该地区的土壤水分动态监测和生态恢复具有重要的意义。为掌握延河流域的土壤水分变化和分布状况,利用Landsat8 OLI影像数据获取的归一化植被指数（NDVI）和陆地表面温度（Ts）,构建Ts-NDVI特征空间,分析土壤水分空间分布变化。结合野外148个实测土壤水分数据,建立0~20 cm,20~40 cm,40~60 cm,60~80 cm和80~100 cm深度的土壤水分遥感反演回归模型,对比分析了5个深度土壤水分的空间变化特征,利用剩余30个样点数据进行精度验证。结果表明：延河流域微旱和干旱的分布范围广且多分布于西北地区,各地表覆盖类型的TVDI总体趋势为：低覆盖度 > 中覆盖度 > 高覆盖度。各土层深度的实测的土壤水分与TVDI的反演值具有较好的负相关性,其中20~40 cm土层的土壤水分与TVDI反演值的相关性较强,表明TVDI指数法适用于延河流域土壤水分反演。TVDI能更稳定反映和指示20~40 cm土层的土壤水分状况,但对于提取更深土层深度的土壤水分信息不能满足精度要求。     

NDVI相似性分区下天山地区草地总产草量遥感估算

山区草地总产草量遥感估算是定量评价区域牧业生产力的有效手段。常规总产草量地面观测数据准确性较高,但无法覆盖整个天山山区,且耗时耗力。针对此问题,以新疆天山山区为研究区,选取MODIS/MOD13Q1 250m植被指数(normalized difference vegetation index,NDVI)产品数据,以县(市)为单元,基于巴氏距离定量评价研究区植被指数分布区域相似性以得到有效遥感建模分区,在此分区基础上,结合草地总产草量实测数据,建立研究区植被指数-草地总产草量遥感估算模型。结果显示,1)基于各县(市)2009-2015年7月底至8月初植被生长期多年NDVI均值直方图计算巴氏距离,以巴氏距离d0.5为阈值,研究区被划分为7个遥感建模区;2)各分区内NDVI-草地总产草量数据拟合方程形式不同,有线性、指数、幂指数和多项式回归方程几种形式。总体来看,各分区NDVI-草地总产草量拟合相关系数在0.784~0.836。交叉检验除天山北坡西段-伊犁河谷草原畜牧业区RMSE值在2 951kg·hm-2外,其他分区RMSE值均在266~928kg·hm-2,原因在于伊犁河谷草原畜牧业区实测草地总产量在10 000~30 000kg·hm-2的样点居多,区域草地总产量较其他区域多。     

基于综合分段优势的植被覆盖度模型比较

基于Landsat 8 OLI遥感影像选取9种植被指数,与野外实测植被覆盖度进行相关性分析,采用分段方式选择敏感植被指数,构建回归模型及FCD模型(forest canopy density mapping model,FCD)对渭干河–库车河三角洲绿洲植被覆盖度进行反演。结果表明:1)根据实测数据计算不同植被指数在不同植被覆盖度范围内的变化比例,从而确定0.3和0.7作为研究区植被覆盖度的分段点;2)分段回归模型与综合分段优势的FCD模型建模精度均约为79%,但综合分段优势的FCD模型建模集R2较高,RMSE较小,其验证结果优于分段回归模型,验证集R2为0.832,RMSE为0.154,精度为82.018%。综合分段优势的FCD模型更适合用于研究区总体植被覆盖度反演,可为干旱区植被覆盖度定量监测与生态环境评价提供依据。     

基于综合分段优势的植被覆盖度模型比较

基于Landsat 8 OLI遥感影像选取9种植被指数,与野外实测植被覆盖度进行相关性分析,采用分段方式选择敏感植被指数,构建回归模型及FCD模型(forest canopy density mapping model,FCD)对渭干河–库车河三角洲绿洲植被覆盖度进行反演.结果表明:1)根据实测数据计算不同植被指数在不同植被覆盖度范围内的变化比例,从而确定0.3和0.7作为研究区植被覆盖度的分段点;2)分段回归模型与综合分段优势的FCD模型建模精度均约为79％,但综合分段优势的FCD模型建模集R2较高,RMSE较小,其验证结果优于分段回归模型,验证集R2为0.832,RMSE为0.154,精度为82.018％.综合分段优势的FCD模型更适合用于研究区总体植被覆盖度反演,可为干旱区植被覆盖度定量监测与生态环境评价提供依据.     

大针茅典型草原冠层光合有效辐射各分量变化规律分析

利用光量子仪实测了大针茅典型草原植被类型关键生育期光合有效辐射值,分析了入射PAR、冠层反射PAR、透射PAR、土壤反射PAR与FPAR的变化规律。结果表明,入射与透射PAR呈较标准正弦曲线变化、而冠层反射与土壤反射PAR呈不太平滑的正弦曲线变化,最大值均出现在中午12:30左右;FPAR曲线开口方向与PAR的开口方向相反,呈较标准的余弦曲线变化,最小值出现在中午11:30左右;且日均FPAR值可用一天内的上午9:30或下午14:30的瞬时值来替代;NDVI、LAI与FPAR均呈线性关系,且R~2分别达到0.87、0.7。     

Heartworm in dogs in Canada in 1988

In late November 1988, 1581 small and mixed animal clinics and institutional veterinarians across Canada were sent a questionnaire in order to assess the status of Dirofilaria immitis in Canada in 1988, and 46% of them responded. Veterinarians reported that 181,577 dogs were blood-tested for heartworm disease and 367 dogs were found with D. immitis microfilariae. Another 60 dogs were amicrofilaremic but diagnosed with heartworm disease to give the total number of cases diagnosed in 1988 as 441 (0.24%).     

Heartworm in dogs in Canada in 1985

In late December 1985, 1485 institutional veterinarians and small and mixed animal clinics across Canada were sent a questionnaire in order to assess the status of Dirofilaria immitis in Canada in 1985 and 44% of them responded. Veterinarians reported that 137,300 dogs were blood-tested to check for microfilariae and 1210 dogs were found with heartworm. Another 36 dogs were amicrofilaremic but diagnosed with heartworm disease to give the total number diagnosed in 1985 as 1247 (0.91%).

Heartworm was reported from all provinces except Prince Edward Island, Newfoundland and Saskatchewan but most (1126) of the cases were in Ontario. Southwestern Ontario continued to be the primary focus of the infection in Canada. From Quebec, 91 cases were reported mostly from and around Montreal. From Manitoba, 19 cases were reported from Winnipeg and surrounding areas. Heartworm was found most frequently in companion dogs over three years of age maintained mainly outdoors in rural areas. About 28% of the cases were observed with clinical signs of heartworm disease and 78% had a history of not having left Canada.

     

Scrapie in goats in Cyprus

Ovine scrapie was first recorded in Cyprus in 1985. Subsequently four dairy goats kept in two mixed flocks with affected sheep developed characteristic clinical signs similar to those seen in sheep. Fifteen goats from the two flocks were examined histologically and neurological lesions consistent with a diagnosis of scrapie were found in the four animals and in three others which had subsequently developed early neurological signs. These lesions were similar to those of naturally-affected sheep although neuronal degeneration and vacuolation were more severe in some cases.     

Heartworm in dogs in Canada in 1983

In late December 1983, 2 800 veterinarians across Canada were sent a questionnaire in order to assess the status of heartworm disease in Canada in 1983 and 26% of them responded. Veterinarians reported that 59 504 dogs were blood-tested to check for microfilariae and 771 dogs (1.30% of those tested) were found with Dirofilaria immitis. Heartworm disease was diagnosed in all provinces except New Brunswick and Newfoundland but most (733) of the cases were in Ontario.

Heartworm disease was found most frequently in companion dogs over three years of age maintained mainly outdoors in rural areas. About 31% of the cases were observed with clinical signs of heartworm disease and 64% had a history of not having left Canada. Southwestern Ontario continues to be the focus of the infection and most of the dogs there had not left the province previously.

     

Heartworm in dogs in Canada in 1984

In late December 1984, 1853 institutional veterinarians and small and mixed animal clinics across Canada were sent a questionnaire in order to assess the status of Dirofilaria immitis in Canada in 1984 and 35% of them responded. Veterinarians reported that 97,794 dogs were blood-tested to check for microfilariae and 1417 dogs (1.45% of those tested) were found with heartworm. Another 34 dogs were amicrofilaremic, but were diagnosed as having heartworm disease, to give the total number diagnosed in 1984 as 1451 (1.48%). Heartworm was reported from all provinces except Prince Edward Island and Newfoundland but most (1310) of the cases were in Ontario. In Quebec, 126 cases were reported mostly from west of Montreal.

Heartworm was found most frequently in companion dogs over three years of age maintained mainly outdoors in rural areas. About 27% of the cases were observed with clinical signs of heartworm disease and 72% had a history of not having left Canada. Southwestern Ontario continued to be the primary focus of the infection.

     

Heartworm in dogs in Canada in 1989

In late November 1989, 1732 clinics and institutional veterinarians were sent a questionnaire to assess the status of Dirofilaria immitis, and 51.7% responded. Of 247,716 dogs tested, 394 had D. immitis microfilariae and 51 were amicrofilaremic for a total of 445 cases and heartworm prevalence of 0.17%. Most (408) of these dogs had no preventive medication and the prevalence among dogs tested and unprotected was 1.01%. That prevalence was considerably higher in endemic areas. Thirty-seven dogs with heartworm had preventive medication. Heartworm was most frequent in companion dogs over three years of age maintained outdoors in rural areas. About 75% of the cases had never left Canada, 26% had clinical signs and 125 were not treated.

Heartworm was reported from British Columbia, Manitoba, Ontario, Quebec, Nova Scotia and Newfoundland, but 383 cases were in Ontario. South-western Ontario was the primary focus of infection. There were 33 cases in Quebec and 24 in Manitoba, mainly found in and around Metropolitan Montreal and Winnipeg respectively.

     

Heartworm in dogs in Canada in 1987

In late November 1987, 1246 institutional veterinarians and small and mixed animal clinics across Canada were sent a questionnaire in order to assess the status of Dirofilaria immitis in Canada in 1987, and 50% of them responded. Veterinarians reported that 165,428 dogs were blood tested for heartworm disease and 511 dogs were found with D. immitis microfilariae. Another 78 dogs were amicrofilaremic but diagnosed with heartworm disease to give the total number of cases diagnosed in 1987 as 589 (0.35%).