Estimation of Vegetation Water Content from Landsat 8 OLI Data
ZHENG Xing-ming1, 2, DING Yan-ling1, ZHAO Kai1, 2*, JIANG Tao1, LI Xiao-feng1, 2, ZHANG Shi-yi1, LI Yang-yang1, WU Li-li1, SUN Jian3, REN Jian-hua1, ZHANG Xuan-xuan4
1. Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102, China 2. Changchun Jingyuetan Remote Sensing Test Site, Chinese Academy of Sciences, Changchun 130102, China 3. College of Electronic Science and Engineering, Jilin University, Changchun 130012, China 4. Institute for Geo-Informatics & Digital Mine Research, Northeastern University, Shenyang 110819, China
摘要： 植被含水量是作物长势好坏的指示因子，利用遥感技术及时准确监测植被含水量对农业生产、作物估产和干旱状况评价具有重要意义。基于新一代对地观测计划Landsat 8 OLI传感器(Operational Land Imager，陆地成像仪)，评价其植被含水量反演的能力与局限性。首先，利用ProSail冠层模型模拟冠层光谱反射率数据集，分析OLI传感器的植被含水量敏感波段以及土壤背景对各波段反射率的影响，然后利用基于Landsat OLI影像计算的植被水分指数和2013年6月1日—8月14日期间采样的植被含水量数据，比较12种植被水分指数与地面实际采样的植被含水量的相关性，评价估算植被含水量的最佳植被水分指数。结果表明：OLI传感器的红、近红外和两个短波红外对植被含水量敏感，其中近红外波段最为敏感;在低植被覆盖度时，土壤背景反射率的太阳辐射将达到光谱传感器影响植被水分指数与植被含水量之间的关系，利用ProSail模拟干湿土壤背景反射率结果也表明土壤背景对植被冠层反射率的影响很大;引入优化土壤调整植被指数(OSAVI)去除土壤背景对植被水分指数的影响;在12种植被水分指数中，MSI2与植被含水量的拟合关系最好(R2=0.948)，植被含水量的平均拟合误差为0.52 kg·m-2;在植被生长晚期即植被含水量大于2 kg·m-2时，各植被水分指数出现饱和情况，植被含水量的估算结果不佳。
关键词：植被含水量;Landsat 8;植被水分指数;ProSail模型
Abstract：The present paper aims to analyze the capabilities and limitations for retrieving vegetation water content from Landsat8 OLI (Operational Land Imager) sensor-new generation of earth observation program. First, the effect of soil background on canopy reflectance and the sensitive band to vegetation water content were analyzed based on simulated dataset from ProSail model. Then, based on vegetation water indices from Landsat8 OLI and field vegetation water content during June 1 2013 to August 14 2013, the best vegetation water index for estimating vegetation water content was found through comparing 12 different indices. The results show that: (1) red, near infrared and two shortwave infrared bands of OLI sensor are sensitive to the change in vegetation water content, and near infrared band is the most sensitive one; (2) At low vegetation coverage, solar radiation reflected by soil background will reach to spectral sensor and influence the relationship between vegetation water index and vegetation water content, and simulation results from ProSail model also show that soil background reflectance has a significant impact on vegetation canopy reflectance in both wet and dry soil conditions, so the optimized soil adjusted vegetation index (OSAVI) was used in this paper to remove the effect of soil background on vegetation water index and improve its relationship with vegetation water content; (3) for the 12 vegetation water indices, the relationship between MSI2 and vegetation water content is the best with the R-square of 0.948 and the average error of vegetation water content is 0.52 kg·m-2; (4) it is difficult to estimate vegetation water content from vegetation water indices when vegetation water content is larger than 2 kg·m-2 due to spectral saturation of these indices.
Key words：Vegetation water content;Landsat8;Vegetation water index;ProSail model
[1] Ceccato P, Gobron N, Flasse S，et al. Remote Sens. Environ., 2002, 82: 188. [2] Zarco-Tejada P J, Rueda C A, Ustin S L. Remote Sens. Environ., 2003, 85: 109. [3] ZHANG Jia-hua, XU Yun, YAO Feng-mei, et al(张佳华，许 云，姚凤梅，等). Sci. China Tech. Sci.(中国科学：技术科学)，2010，40(10)：1121. [4] Palmer K F, Williams D. Journal of the Optical Society of America, 1974, 64: 1107. [5] Tucker C. J. Remote Sens. Environ., 1980, 10: 23. [6] Saleem Ullaha, Andrew K Skidmorea, et al. Agricultural and Forest Meteorology, 2013, 171: 65. [7] Zhang J H, Guo W J. Proc. SPIE 6411, Agriculture and Hydrology Application of Remote Sensing, 2006, 64110D, DOI: 10.1117/12.697957. [8] Chuvieco E, Deshayes M, Stach N, et al. Remote Sensing of Large Wildfires: in the European Mediterranean Basin, Springer-Verlag, Berlin, 1999. 7. [9] Danson F M, Steven M D, Malthus T J, et al. Int. J. Remote Sens., 1992, 13(3): 461. [10] Knipling E B. Remote Sens. Environ., 1970, 1: 155. [11] Carter G A. Am. J. Bot., 1991, 78: 919. [12] Kou L, Labrie D, Chylek P. Appl. Opt., 1993, 32: 3531. [13] Curran P J. Remote Sens. Environ., 1989, 30: 271. [14] Ceccato P, Flasse S, Tarantola S, et al. Remote Sens. Environ., 2001, 77: 22. [15] Morisette J T, Baret F, Privette J L, et al. IEEE Transactions on Geosciences and Remote Sensing, 2006, 44(7): 1. [16] Dawson T P, Curran P J, North P R J，et al. Remote Sens. Environ., 1999, 67: 147. [17] Danson F M, Bowyer P. Remote Sens. Environ., 2004, 92: 309. [18] Jordan C F. Ecology, 1969, 50: 663. [19] Trombetti M, Riao D, Rubio M A，et al. Remote Sens. Environ., 2008, 112: 203. [20] Fensholt R, Sandholt I. Remote Sens. Environ., 2003, 87: 111. [21] ZHANG You-shui, XIE Yuan-li(张友水, 谢元礼). Scientia Geographica Sinica(地理科学), 2008, 28(1): 72. [22] WANG Zheng-xing, LIU Chuang, Huete A(王正兴，刘 闯，Huete A). Acta Ecologica Sinica(生态学报) 2003，23(5)：979. [23] Serrano L, Ustin S L, Roberts D A, et al. Remote Sens. Environ., 2000, 74: 570. [24] Hardinsky M A, Lemas V, Smart R M. Photo Eng. Remote Sen., 1983, 49: 77. [25] Chen D, Huang J F, Jackson T J. Remote Sens. Environ., 2005, 98: 222. [26] Gao B C. Remote Sens. Environ., 1996, 58: 257. [27] Datt B. Australian Journal of Botany, 1999, 47: 909. [28] Wang L, Qu J J. Geophys. Res. Lett., 2007, 34: L20405. [29] Huete A, Didan K, Miura T, et al. Remote Sens. Environ., 2002, 83: 195. [30] Rondeaux G, Steven M, Baret F. Remote Sens. Environ., 1996, 55(2): 95
