The vegetation greenness trend in Canada and US Alaska from 1984-2012 Landsat data
The vegetation greenness trend in Canada and US Alaska from 1984-2012 Landsat data
复制标题
DOI：
10.1016/j.rse.2016.01.001
复制
发表时间：
2016-04-01
影响因子：
13.5
通讯作者：
Masek, Jeffrey G.
中科院分区：
文献类型：
--
作者：
Ju, Junchang;Masek, Jeffrey G.
为了评估北美高纬度植被对气温上升的响应，我们使用1984-2012年夏季Landsat地表反射率数据，推导了加拿大和阿拉斯加91.2%的无水无雪土地面积的NDVI趋势。我们的分析表明，29.4%和2.9%的土地面积的加拿大和阿拉斯加分别表现出统计上显着的积极（绿化）和消极（布朗宁）的趋势，在显着性水平p < 0.01，后烧毁的森林面积被掩盖。在所有土地覆被类型中，绿化趋势区的面积均大于布朗宁趋势区。绿化主要发生在阿拉斯加西部的苔原，沿着加拿大北部海岸和加拿大东北部;最密集和广泛的绿化发生在魁北克和拉布拉多。布朗宁主要发生在阿拉斯加东部的北方森林。对于所有植被带，基于陆地卫星的绿色趋势与基于8公里GIMMS AVHRR的趋势大致相似。然而，对于苔原，Landsat数据表明，阿拉斯加北坡的绿化范围要小得多，魁北克和拉布拉多的绿化范围要大得多，而北方森林的布朗宁趋势要小得多，这些森林没有受到火灾的干扰。这些差异要求进一步验证大地卫星反射率和AVHRR NDVI数据集。需要与地形、冰川历史和土壤条件等当地环境因素进行相关研究，以了解陆地卫星尺度上的非均匀绿色变化。(C)2016 Elsevier Inc. All rights reserved.
To assess the North American high-latitude vegetation response to the rising temperature, we derived NDVI trend for 91.2% of the non-water, non-snow land area of Canada and Alaska using the peak-summer Landsat surface reflectance data of 1984-2012. Our analysis indicated that 29.4% and 2.9% of the land area of Canada and Alaska showed statistically significant positive (greening) and negative (browning) trends respectively, at significance level p < 0.01, after burned forest areas were masked out. The area with greening trend dominated over that with browning trend for all land cover types. The greening occurred primarily in the tundra of western Alaska, along the north coast of Canada and in northeastern Canada; the most intensive and extensive greening occurred in Quebec and Labrador. The browning occurred mostly in the boreal forests of eastern Alaska. The Landsat-based greenness trend is broadly similar to the 8-km GIMMS AVHRR-based trend for all vegetation zones. However, for tundra, the Landsat data indicated much less extensive greening in Alaska North Slope and much more extensive greening in Quebec and Labrador, and substantially less extensive browning trend in the boreal forests that were free of fire disturbances. These differences call for further validation of the Landsat reflectance and the AVHRR NDVI datasets. Correlation study with local environmental factors, such as topography, glacial history and soil condition, will be needed to understand the heterogeneous greenness change at the Landsat scale. (C) 2016 Elsevier Inc. All rights reserved.