基于遥感和气象数据的江苏省水稻面积监测和估产研究

发布时间：2018-05-24 21:55

  本文选题：遥感 + 水稻　； 参考：《浙江大学》2016年硕士论文

【摘要】：“国以民为本,民以食为天”,在当今风云突变的气候条件和经济形势下,中国及全球主要产粮区粮食作物生长状况监测和产量估算等措施的及时有效执行,密切关乎到粮食安全问题。我国作为世界第一人口大国,水稻第一生产和消费大国,为针对水稻的生长状况的监测研究提出了高要求。在此背景之下本研究以江苏省水稻遥感监测作为研究目标,利用遥感数据在客观及时性、宏观尺度、经济效益等方面上的优势,提取水稻面积,估算水稻产量,为更合理制定农业政策、宏观调控粮食价格、缓解粮食危机提供可能。本文利用时间序列MOD09A1数据提取江苏2004-2010年的水稻面积,并利用中等空间分辨率Landsat 8 OLI数据提取江苏省水稻种植区域,以其为准确值,分析基于低空间分辨率MOD09A1数据江苏省水稻播种区域识别结果空间匹配状况的准确性。同时以2004-2006三年水稻生育期EVI与产量统计数据构建估产模型,估算江苏省2007年各地级市水稻单产,并将市域尺度的水稻单产转换到单个象元点尺度上,实现水稻单产空间化；基于改进的两叶直角双曲线模型,利用气象数据最低气温、最高气温、相对湿度和向下短波辐射,结合MOD15A2叶面积指数数据,计算得到水稻生育期GPP值,进而估算江苏省水稻产量。本文的主要成果和创新点：1)时间序列MODIS数据提取的水稻面积,在空间分布上有一定准确度,但总体面积偏低；修改水稻提取算法参数,使MODIS数据水稻面积提取与统计数据持平,空间匹配上存在较大误差,特别是在南京和启东等地,但主要产量区都有所提取。2)水稻估产结果在地级市尺度上,精确度维持在95%左右,在此基础上,将水稻单产值转化到单象元尺度上,细化了水稻估产结果,可以更直观的了解稻田生产力信息。3)增加25℃最大羧化速率参数(Vcmax,25)以构建了日尺度两叶直角双曲线模型估算每日GPP,并用于区域水稻估产。
[Abstract]:"the country is based on the people and the people take food as the sky." under the sudden climatic conditions and economic situation, the measures such as monitoring the growth status of grain crops and estimating the output of grain crops in major grain producing areas in China and the world are implemented in a timely and effective manner. It is closely related to food security. As the largest population country in the world and the largest producer and consumer of rice in the world, China has put forward high requirements for monitoring and research on the growth of rice. Under this background, this research takes the rice remote sensing monitoring in Jiangsu Province as the research goal, utilizes the remote sensing data in the objective timeliness, the macroscopic scale, the economic benefit and so on superiority, extracts the rice area, estimates the rice yield, etc. For a more rational formulation of agricultural policies, macro-control of grain prices, to ease the food crisis possible. The rice area of Jiangsu Province from 2004 to 2010 was extracted by time series MOD09A1 data, and the area of rice cultivation in Jiangsu Province was extracted by Landsat 8 OLI data with medium spatial resolution, which was regarded as the accurate value. Based on low spatial resolution MOD09A1 data, the accuracy of spatial matching of rice planting area recognition results in Jiangsu Province was analyzed. At the same time, based on the EVI and yield statistical data of three years of rice growth period from 2004 to 2006, the yield estimation model was constructed, and the rice yield per unit yield in Jiangsu Province in 2007 was estimated, and the rice yield in city scale was converted to a single pixel point scale to realize the spatialization of rice yield per unit yield. Based on the improved two-leaf right-angle hyperbolic model, the GPP value of rice growing period was calculated by using meteorological data, such as minimum temperature, maximum temperature, relative humidity and downward shortwave radiation, combined with MOD15A2 leaf area index data. And then estimate the yield of rice in Jiangsu Province. The main achievements and innovations of this paper are: 1) the rice area extracted by time series MODIS data has some accuracy in spatial distribution, but the total area is on the low side, and the parameters of rice extraction algorithm are modified to make the rice area extraction from MODIS data equal to the statistical data. There is a large error in spatial matching, especially in Nanjing and Qidong, but the main yield areas have been extracted. 2) the precision of rice yield estimation is about 95% on the municipal scale. On this basis, The single output value of rice was transformed to single pixel scale, and the estimated yield of rice was refined. The maximum carboxylation rate was increased by 25 鈩，

本文编号：1930723