林业科学  2015, Vol. 51 Issue (2): 90-98 PDF
DOI: 10.11707/j.1001-7488.20150211
0

#### 文章信息

Miao Qinglin, Tian Xiaorui, Zhao Fengjun

NDVI Recovery Process for Post-Fire Vegetation in Daxing'anling

Scientia Silvae Sinicae, 2015, 51(2): 90-98.
DOI: 10.11707/j.1001-7488.20150211

### 作者相关文章

NDVI Recovery Process for Post-Fire Vegetation in Daxing'anling
Miao Qinglin, Tian Xiaorui, Zhao Fengjun
Key Laboratory of Forest Protection of State Forestry Administration Research Institute of Forest Ecology, Environment and Protection, CAF Beijing 100091
Abstract: [Objective]The remote sensing technology was used to monitor the vegetation restoration after fire, providing a scientific base for carrying out restoration measures in burned areas. The normalized difference vegetation index (NDVI) is an important index to reflect the growth condition and distribution of vegetation. It has been proved in previous studies that this index has a significant correlation with vegetation coverage. Thus the increasing biomass and the vegetation coverage in burned areas can be monitored through the satellite remote sensing images.[Method] The Songling burned area, which was burned in spring of 2006, in Daxing'anling was selected as a case study. A series of NDVI data before and after the fire, which were extracted from the MODIS data, and the field investigation data were used to analyze the relationships between vegetation characteristics after fire, burned severity and vegetation types. Data of NDVI in the burned area were extracted before and just after the fire, and the fire severity was classified using the supervised classification method. The maximum NDVI in the same date of August in 2003-2005 was used as the contrast to analyze the vegetation index changes on the time series. [Result]Low, moderate and high burning severities were accounted for 28.93%, 40.1% and 30.97% burned area, respectively. The dominated vegetation types with high-burning severity were evergreen coniferous forest, broadleaf and conifer mixed forest, and brushwood, which were accounted for 50.37%, 52.22%, and 59.49%, respectively. The proportion of high severity burned areas increased with the ascending slope. [Conclusion] The post-fire NDVI showed a increasing trend generally. NDVI value of each vegetation type in the area with high-burning severity was significantly lower than the low and moderate burning severity areas, except for the grassland. But there was no significant difference in NDVI between the areas with low and moderate burning severity. In the second year, the vegetation coverage in high burning severity areas reached the minimum. The NDVI of these vegetation types in low burning severity areas recovered to pre-fire level in 6 years after fire. The coverage of broadleaf and conifer mixed forests recovered faster than other forest types. Fire severity affected forest vertical structure. The burned forests had greater shrub coverage than un-burned ones, and this phenomenon was more obvious in the forests with high fire severity. The natural restoration of brushwood, grassland and marsh was faster than that of forests, thus these areas don't need artificial aids to update. Natural restoration of the tree layer in forests with high-burning severity is very slow, the artificial update will speed up the succession process of forest communities. Periodic drought has an influence on NDVI, especially for the post-fire grassland. The two-factor ANOVA showed that vegetation type and fire severity had a significant influence on the vegetation index. dNDVI can reflect the changes of the vegetation well, which has a good temporal and spatial availability and plays an important role in monitoring the post-fire vegetation restoration.
Key words: burned area    NDVI    vegetation restoration

1 研究区概况

2 数据来源

3 研究方法 3.1 过火区提取

 ${\text{NDVI}} = \left({{R_{{\text{nir}}}} - {R_{{\text{red}}}}} \right)/\left({{R_{{\text{nir}}}} + {R_{{\text{red}}}}} \right)$ (1)

3.2 火烧强度等级分类

 ${\text{dNDVI}} = {\text{NDV}}{{\text{I}}_{{\text{post，t}}}} - {\text{NDV}}{{\text{I}}_{{\text{pre，m}}}}$ (2)

3.3 标准地调查

2012年6月在研究区内设置16块30 m×30 m的标准地，包括火烧区5种主要植被类型，即落叶针叶林、针阔混交林、落叶阔叶林、灌丛和草地(其中灌丛和草地只设重度火烧和未火烧区)，分别调查未火烧、轻度火烧、中度火烧和重度火烧情况下的植被状况。调查指标包括郁闭度、平均胸径、平均树高、平均枝下高、死亡率、平均熏黑高度、下木盖度、下木平均高度、草本盖度和草本高度。

3.4 可燃物湿度码

4 结果与分析 4.1 过火区的植被类型与火烧强度等级

 图 1 火烧强度等级分类 Fig. 1 Burned rating classification
 图 2 不同坡度上各火烧强度等级比例 Fig. 2 Percentage for burned rating at different slopes

 图 3 火烧强度等级-植被类型图 Fig. 3 Map of burned rating and vegetation type
4.2 火后植被的NDVI变化

 图 4 不同植被类型与火烧强度的dNDVI变化 Fig. 4 The dNDVI curves for different vegetation types V1：常绿针叶林Evergreen coniferous forest； V2：落叶针叶林Deciduous coniferous forest； V3：落叶阔叶林Deciduous broadleaved forest； V4：针阔混交林Broadleaved and coniferous mixed forest； V5：灌丛Brushwood； V6：草地Grassland； V7：沼泽Marsh； A：过火区平均值Average for the burned area.

4.3 火后植被NDVI的影响因子

NDVI与植被含水量存在一定相关性，并受很多因子影响，包括温度、降水、空气相对湿度、立地条件、植被类型等。方差分析结果(表 2)表明，火烧强度等级和植被类型对火后植被NDVI都有显著影响(在ɑ=0.01水平上显著)，但二者没有交互作用。

 图 5 2006—2013年过火区与未过火区dNDVI对照 Fig. 5 Comparative analysis of the dNDVI for the burned and unburned area from 2006 to 2013

NDVI对空气温度和降水等的响应具有滞后性，且夏季滞后期更长(崔林丽等，2009)。主要植被类型NDVI在2007和2010年出现较低值，与阶段性干旱有一定关系。这2年夏季降水量偏低，6—7月降水量距平分别为41.81和66.81 mm(图 6)。DMC的时段(06-01—08-13)平均值也明显偏高，2007和2010年分别为55.20和55.68(图 7)，阶段最高值分别为78.56和94.71。干旱对NDVI有显著影响，导致NDVI出现较大程度降低。2013年黑龙江汛期平均降雨量较常年偏多3~4成，并出现超50年一遇特大洪水(那济海等，2013)，超强降水和洪涝灾害可能使2013年植被生长受到影响，2013年NDVI较2012年略有降低。

 图 6 2006—2013年6—7月降水量与气温 Fig. 6 The precipitation and average temperature for June-July in 2006-2013
 图 7 2006—2013年06-01—08-13日均DMC Fig. 7 The average DMC for the period from June 1 to August 13 in 2006-2013
5 结论与讨论

NDVI差值表示火烧前后植被指数变化，能很好地反映过火区火后NDVI的整体趋势，有较好的时序性和空间可获取性，对火烧迹地恢复研究具有重要辅助作用。另外，NDVI差值方法可以作为分析影响火后NDVI变化因子的重要手段。MODIS-NDVI遥感数据为在长时间和大空间尺度上进行火烧迹地植被恢复过程研究与连续监测提供了重要手段；但要注意高覆盖度植被区的NDVI在生长旺盛期易出现饱和现象(李红军等，2007)，即NDVI对高植被区的灵敏度较低，其对植被恢复特征表征准确性的影响有待进一步研究。