-
植被是生态系统的重要组成部分,在陆地表面能量交换过程、生物地球化学循环过程和水文循环过程中扮演着重要的角色[1-3]。传统的植被覆盖度地面测量方法有目估法、采样法、仪器法和照相法等[4]。随着遥感技术的发展,植被指数在植被覆盖度测量上得到了广泛应用[5]。归一化植被指数(Normalized Difference Vegetation Index, NDVI)为遥感影像中近红外波段反射值与红光波段反射值之差比上两者之和,是目前评估植被状况最有效的参数之一。NDVI能够反映出植物冠层的背景影响,范围为−1~1,负值表示地面覆盖为云、水、雪等;0表示有岩石或裸土等;正值表示有植被覆盖,且随覆盖度的增大而增大[6]。国内外学者利用NDVI进行了一系列的植被生态环境研究[7-14]。TANG等[15]采用无人机遥感技术对中分辨率成像光谱仪(Moderate-resolution Imaging Spectroradiometer, MODIS)NDVI数据产品植被覆盖度的反演结果进行了验证,结果表明,利用NDVI数据产品反演沙漠地区植被覆盖率是可行的,同时对NDVI数据产品进行校正可以有效提高反演精度。JAFARI等[16]评估了MODIS增强植被指数(Enhanced Vegetation Index, EVI)和NDVI在伊朗中部干旱和半干旱牧场地区的监测能力,发现NDVI更适用于草地状况评价。LONG[17]建立了植被覆盖二分像元反演模型,并用野外调查数据对模型精度进行验证,开展了鸭池示范区植被覆盖度的定量估算研究。王婷[18]基于MODIS NDVI数据反演黑龙江兰西县植被覆盖度,结果表明,NDVI的估算精度可达80%。何国兴等[19]以甘肃省高寒草甸为研究区,基于2000—2019年遥感数据和2014年实测数据,采用经验回归模型法构建植被覆盖度估算模型,结果表明,高寒草甸植被覆盖度的最佳反演植被指数为NDVI。刘晓菲[20]研发了1种区域生长研究区的NDVI提取方法,提取出了江苏省金坛市植被覆盖度并进行验证,发现NDVI有较好的反演效果。李丽丽等[21]基于MODIS NDVI数据利用像元二分模型反演植被覆盖度,分析了2000—2015年石羊河流域植被时空动态变化趋势。李晓松等[22]利用NDVI对荒漠化地区的植被覆盖度进行了估测,其研究结果表明,NDVI能够有效地估测荒漠化地区的植被覆盖度。臧淑英等[23]通过获取大庆1988、1996、2001年3个时段的TM影像,使用NDVI和SAVI两种植被指数,采用像元二分模型法和经验型法建立了大庆地区草地植被盖度的遥感定量反演模型,该模型的精度达到70.34%。
海南岛是我国唯一的热带气候岛屿,在充沛的光、热条件下,植被覆盖茂密丰富,确定合适的技术方法表征海南岛植被覆盖面积时空变化特征是值得研究的问题。前人的研究未涉及海南岛植被覆盖面积反演,基于此,笔者采用一元线性拟合、均方根误差分析、相关系数分析等方法开展基于MODIS NDVI的海南岛植被覆盖面积反演技术研究,旨在为合理监测海南岛植被生态环境变化提供理论依据。
Research on the inversion technique of vegetation coverage in Hainan island based on MODIS NDVI
doi: 10.15886/j.cnki.rdswxb.2022.04.011
- Received Date: 2021-09-15
- Accepted Date: 2022-04-01
- Rev Recd Date: 2022-03-09
- Available Online: 2022-04-08
- Publish Date: 2022-07-26
-
Key words:
- Hainan Island /
- NDVI /
- inversion of vegetation coverage
Abstract: In order to provide theoretical basis for reasonable monitoring of vegetation ecological environment changes in Hainan Island, based on MODIS NDVI data and land cover data of Hainan Island in 2015 and 2017, the inversion technology of vegetation cover area in Hainan Island was studied by univariate linear fitting, root mean square error analysis and correlation coefficient analysis. The results indicated that the main type of land cover in Hainan Island was forest in 2015 which was mainly distributed in Wuzhishan mountains; Secondly, cultivated land was mainly distributed in Haikou, Lingao, Chengmai, Wenchang, Dongfang and other coastal areas; The vegetation coverage area (forest, grassland, shrub and wetland) accounted for 65.67% of the total area. The surface coverage in 2017 was similar to that in 2015. The vegetation coverage of Hainan Island is generally high, and the vegetation coverage of cities in the central mountainous area is higher than that of coastal cities. Wuzhishan has the highest vegetation coverage rate, followed by Qiongzhong and Baisha; Haikou had the lowest vegetation coverage, followed by Wenchang and Ding'an. The best threshold of NDVI vegetation coverage inversion in Hainan Island is 0.70, under this threshold, the spatial recognition rates of NDVI for vegetation cover in 2015 and 2017 reached 78.64% and 78.74% respectively; The best fitting regression model of vegetation coverage area and NDVI inversion value in Hainan Island was y=0.8988x+ 62708142.60, the correlation coefficient is 0.91 and the root mean square error accounted for 17.50% of the actual mean value, the error is reduced by 7.23% compared with the inversion value. The fitting deviation rate of vegetation coverage area in coastal cities was larger than that in inland cities.
Citation: | YANG Jing, ZHANG Yajie, ZHANG Jinghong, ZHANG Mingjie. Research on the inversion technique of vegetation coverage in Hainan island based on MODIS NDVI[J]. Journal of Tropical Biology, 2022, 13(4): 397-403. doi: 10.15886/j.cnki.rdswxb.2022.04.011 |