【作者】 王冰；

【导师】 黄华国；

【作者基本信息】 北京林业大学 ， 森林经理学， 2019， 硕士

【摘要】 云南地区小蠹虫虫害较为严害,造成了云南松的大量死亡,严重影响当地的生态平衡。实现虫害的早期识别极为重要,利用热红外遥感实现森林虫害的早期监测是近年来的研究热点。地表温度(Land surface temperature,LST)作为基本的物理量,与各项生命活动紧密相关。近些年来,LST在森林病虫害监测方面也有较多的应用。本研究主要尝试探索一种以温度为变量,建立起的基于温度异常的云南松虫害预警体系。本文首先通过定点连续拍摄的红外相机热图像分析了不同受害程度的林地的温度日变化情况,讨论了地表温度的分布规律;然后,通过虫害爆发前后连续多年的Landsat温度产品分析了健康林地与受害林地温度等级分布的差异性以及健康林地与受害林地之间的温差在年际间的变化规律,并利用ENVI-met小气候模型,以样地实测数据和NCEP气象数据为数据源,验证了林分特征参数与地表温度热异常的关系;最后,尝试建立了高分数据中红外波段DN值与Landsat数据反演得到的地表温度产品(LST)的相关关系,为获取高时空分辨率的温度相关产品提供一种理论参考。论文的主要结论如下:1)对不同受害程度的样地的组分温度日变化的分析发现,无论是林木温度还是地表温度,从9时到13时左右,就健康样地而言,光照面温度上升的幅度呈现不规律性变化,而阴影面温度上升的幅度呈现一种逐步增大的趋势,相对于健康样地而言,受害样地的林木温度与地表温度的自我调节能力较差,温度持续性增大或减小,并且受害程度越严重,林分的自我调节能力越差;利用红外热图像获取地表温度信息受到林下地表覆盖类型的影响;12时左右可能是受害样地与健康样地温度分布规律差异最大的时间段,更有利于林分受害程度的区分。2)地表温度等级分布差异性分析发现,旱季受害林地的地表温度平均值低于健康林地,在2013年虫害大爆发之前温差不断增大,在虫害大爆发之后温差开始有所下降并逐渐趋于稳定;2009年严重干旱爆发之后,受害林地与健康林地之间的温度等级分布差异性进一步增大,随后渐趋于稳定;从2012年起,所处温度等级较高的像元百分比有下降,所处温度等级较低的像元百分比有所上升,但是,相对于健康林分,受害林分中,温度等级较低的像元所占百分比上升幅度更大;在林分特征参数与地表温度热异常的相关关系的研究中,树高、LAI、株树密度等参数与地表温度热异常之间可能存在着相关性,这些参数可能主要是通过影响土壤组分温度的分布来对地表温度分布产生影响。3)高分四号中红外波段DN值与Landsat数据反演得到的温度产品呈线性相关,相关系数为0.66。这一尝试可为高时空分辨率数据的生成提供理论参考。更多还原

【Abstract】 The area of forest land affected by pests in China is increasing,Tomicus yunnanensis and Tomicus minor have caused serious shoot damage in Yunnan pine forests in the Yunnan Province of China,causing a large number of deaths of Yunnan pine,which restricts the local ecological balance.It is extremely important to realize the early identification of pests.Early detection of forest pests using thermal infrar-ed remote sensing is a research hotspot in recent years.Land surface temperature(LST)can directly est-imate long-wave radiant flux,and can also estimate the latent heat and sensible heat flux affecting the atmosphere,ocean and land indirectly,it is the basic physical quantity for studying the radiation and energy balance of the gas system and the interaction of the earth and air.LST are widely used in forest pest monitoring.This study mainly attempts to explore an early warning system for pests based on tem-perature anomalies in Yunnan.In this paper,the daily variation of temperature of forest land with different degrees of damage is analyzed by the thermal image taken continuously at a fixed position,the study also discussed the distribution of surface temperature;Then,through the Landsat temperature products for many years before and after the pest outbreak,the differences in temperature distribution between healthy forest land and the affected forest land and the temperature difference between healthy forest land and the affected forest land were analyzed.Based on the measured data and NCEP meteorological data,the relationship between the stand parameters and the surface temperature thermal anomaly was verified by the ENVI-met microclimate model;Finally,the research attempts to establish the correlation between the DN value of GF-4 in the mid-infrared band and the surface temperature product(LST)obtained by Landsat datas(level 1),which provides a theoretical reference for obtaining temperature-related products with high spatial and temporal resolution.The main conclusions of the paper are as follows:1)Analysis of the daily variation of the component temperature of the plots with different degrees of damage found that,regardless of the forest temperature or the surface temperature,from 9 to 13 o’clock,the temperature rise of the healthy plots is irregularly changed.,the temperature rise of the shadow surface of the healthy sample gradually increases;Compared with healthy plots,the self-regulation ability of forest temperature and surface temperature of the affected plot is poor,the temperature persistence increases or decreases,and the more serious the damage,the worse the self-regulation ability of the stand;In addition,the use of infrared thermal images to obtain surface temperature information is affected by the type of surface coverage under the forest.2)The analysis of the difference in the distribution of surface temperature grades shows that the average surface temperature of the affected forest land in the dry season is lower than that in the healthy forest land.Before the outbreak of pests in 2013,the temperature difference increased continuously,after the outbreak of pests,the temperature difference began to decrease and gra-dually stabilized;After the severe drought in 2009,the difference in temperature distribution between the affected forest land and the healthy forest land further increased,and then gradually stabilized;Since 2012,the percentage of pixels with higher temperature grades has decreased,and the percentage of pixels with lower temperature grades has increased.However,compared with healthy forest stands,the percentage of pixels with lower temperature grades increased more in the affected forests;In the study of the correlation between stand parameters and surface temperature thermal anomalies,there may be a correlation between tree height,LAI,plant density and other forest stand parameters and surface temperature thermal anomalies.These parameters may mainly affect the surface temperature distribution by affecting the distribution of soil component temperature;At around 12:00,it may be the time period in which the temperature distribution of the victim plot and the health sample are the most different.3)The DN value of the mid-infrared band of GF-4 is linearly related to the temperature product obtained by the Landsat data inversion,with a correlation coefficient of 0.66.This attempt provides a theoretical reference for the generation of temperature dependent parameters with high spatial and temporal resolution.更多还原