科研圈内人 2022-04-27 21:00
第70期
在国家尺度实现所有森林和非森林树木的数字孪生
题目：Towards digital twinning of all forest and non-forest trees at national levels (发表于Nature，并入选 Nature 2020年度十大科学发现)
参会方式：线上参会（具体参会方式请扫描文末二维码报名选择）
会议时间：报告总时长约35分钟, 留言板&语音提问讨论25~40分钟
北京时间 4月29日（周五）晚上9:00 （欧洲专场特别时间）
纽约时间 4月29日（周五）上午9:00
哥本哈根时间 4月29日（周五）下午3:00
图片
报告人简介
Dr. Martin Brandt received his PhD from the University of Bayreuth, Germany in 2014. He works as Assistant Professor at the University of Copenhagen since 2015. His major fields are remote sensing and physical Geography and his work concentrates on the monitoring of vegetation dynamics, with a special focus on West Africa and Southern China. Martin Brandt’s recent work was the first applying deep learning on hundreds of thousands of sub-metre satellite images to map billions of individual tree crowns in desert areas that were often assumed as free of any trees.(https://www.treesoutsideforests.com/)
Martin Brandt教授，于2014年获得德国拜罗伊特大学博士，2015年担任哥本哈根助理教授。其主要研究领域为遥感和自然地理学，从事西非和中国南方地区植被动态监测研究。最近的工作是首次在亚米级卫星图像上应用深度学习来绘制沙漠地区数十亿棵树冠的地图，这些地区通常被认为没有任何树木。本研究成果入选《Nature》2020年度十大科学发现，实验室主页：https://www.treesoutsideforests.com/
报告摘要
Trees sustain livelihoods and mitigate climate change, but a predominance of trees outside forests and limited resources make it difficult for many countries to conduct frequent nation-wide inventories.The paper will present a rapid and accurate approach to map each individual tree and shrub and their carbon stocks at the national scales using very high resolution satellite and aerial images and deep learning. The mapping of all trees facilitates any landscape stratification and is urgently needed for effective planning and monitoring of landscape restoration activities as well as for optimization of carbon sequestration, biodiversity and economic benefits of trees.
树木具有经济价值并能缓解气候变化带来的影响，但森林以外的树木较多，能力有限使得许多国家难以在全国范围内频繁进行林木资源清查。因此，文中将提出一种快速准确的方法，使用高分辨率卫星和航空图像以及深度学习技术，在国家范围内绘制每棵树木、灌木及其碳储量图。这些图有助于进行景观分层，并且能满足有效规划和监测景观恢复过程以及优化树木的碳固存、生物多样性和经济效益的迫切需要。
论文摘要
A large proportion of dryland trees and shrubs (hereafter referred to collectively as trees) grow in isolation, without canopy closure. These non-forest trees have a crucial role in biodiversity, and provide ecosystem services such as carbon storage, food resources and shelter for humans and animals. However, most public interest relating to trees is devoted to forests, and trees outside of forests are not well-documented. Here we map the crown size of each tree more than 3 m² in size over a land area that spans 1.3 million km² in the West African Sahara, Sahel and sub-humid zone, using submetre-resolution satellite imagery and deep learning. We detected over 1.8 billion individual trees (13.4 trees per hectare), with a median crown size of 12 m², along a rainfall gradient from 0 to 1,000 mm per year. The canopy cover increases from 0.1% (0.7 trees per hectare) in hyper-arid areas, through 1.6% (9.9 trees per hectare) in arid and 5.6% (30.1 trees per hectare) in semi-arid zones, to 13.3% (47 trees per hectare) in sub-humid areas. Although the overall canopy cover is low,the relatively high density of isolated trees challenges prevailing narratives about dryland desertification, and even the desert shows a surprisingly high tree density.Our assessment suggests a way to monitor trees outside of forests globally, and to explore their role in mitigating degradation, climate change and poverty.
大部分旱地树木和灌木（以下统称树木）孤立生长，没有树冠封闭。这些非森林树木在生物多样性中发挥着至关重要的作用，并为人类和动物提供碳储存、食物资源和庇护所等生态系统服务。然而，与树木有关的大多数研究都集中在森林上，而森林之外的树木并没有得到很好的记录。在这里，我们使用亚米分辨率的卫星图像和深度学习，在西非撒哈拉、萨赫勒和半湿润地区的 130 万平方公里的土地上绘制了每棵超过 3 平方米的树冠大小。我们检测到超过 18 亿棵树（每公顷 13.4 棵树），树冠中位数为 12 平方米，降雨梯度为每年 0 到 1,000 毫米。树冠覆盖率从超干旱地区的 0.1%（每公顷 0.7 棵树）增加到干旱地区的 1.6%（每公顷 9.9 棵树）和半干旱地区的 5.6%（每公顷 30.1 棵树），增加到 13.3%（47每公顷树木）在亚湿润地区。尽管整体冠层覆盖率较低，但孤立树木的相对高密度挑战了关于旱地荒漠化的流行说法，甚至沙漠也显示出惊人的高树木密度。我们的评估提出了一种监测全球森林以外树木的方法，并探索它们在减缓退化、气候变化和贫困方面的作用。
研究背景
Trees have long been a central element in environmental science and policy in Africa: threats of deforestation, looming desertification, and ‘stop encroaching deserts’ and ‘plant a tree’ campaigns have been on the front pages of newspapers for decades. Most attention is devoted to forests, which are often defined as areas of more than 25% canopy closure. However, trees from outside of forest areas (non-forest trees) support the livelihoods of a rapidly increasing population through the subsistence use of products such as wood (for construction or fuel), food, fodder and medicinal plants; through the cash income obtained from the sale of products; and through ecological benefits such as protection against hazards (for example, erosion), soil improvement, water and nutrient cycling as well as pollination, which in turn improves agricultural productivity. Moreover, trees inarid biomes are an essential factor for the survival and biodiversity of flora and fauna. Finally, trees in farmlands, savannahs and deserts constitute an important—but very variable— carbon pool, and affect the climate by lowering the albedo, by altering aerodynamic roughness and through transpiration. As non-forest trees are becoming increasingly recognized in environmental initiatives across Africa, there is a growing interest in consistently measuring and monitoring trees outside of forests at the level of single trees.
However, whereas the monitoring of forests has been carried out on a routine basis, attempts to quantify the density of trees outside of forests have been limited to small sample sizes or local field surveys. This is because of the scattered nature of dryland trees, which limits assessments based on commonly available satellite technologies (at a resolution of 10 to 30 m) to the canopy coverage per area, which leaves a blind spot with respect to the number, location and size of isolated trees. The limited attention devoted to the quantification of individual trees in drylands has led to misinterpretations of the extent of canopy cover, and to confusion related to the definition of canopy cover (that is, the characteristics of woody plants included in calculations of ‘coverage’). Products designed to assess global tree cover are poorly designed to quantify tree cover in drylands, which has resulted in the prevailing view that dryland areas such as the Sahara or Sahel are largely free of trees.
树木长期以来一直是非洲的环境科学及政策中关键因素，而森林砍伐的威胁、迫在眉睫的荒漠化和“停止侵占沙漠”和“植树”运动一直在进行几十年来一直是报纸的头版。大多数人关注都集中于树冠闭合率超过25% 的森林。然而，林区以外的树木（非林木）通过使用木材（用于建筑或燃料）、食品、饲料和药用植物等产品维持人口生计；通过生态效益，例如防止危害（例如侵蚀）、土壤改良、水和养分循环以及授粉，这反过来又提高了农业生产力。此外，树木无水生物群落是动植物生存和生物多样性的重要因素。最后，农田、大草原和沙漠中的树木构成了一个重要但变化很大的碳库，并通过降低反照率、改变空气动力学粗糙度和蒸腾作用来影响气候。随着非森林树木越来越多地在非洲的环境倡议中得到认可，人们对在单棵树木的水平上持续测量和监测森林以外的树木越来越感兴趣。
虽然森林监测是在常规基础上进行的，但量化森林外树木密度的尝试仅限于小样本或当地实地调查。这是因为旱地树木的分散性质，这将基于常用卫星技术（分辨率为 10 至 30 m）的评估限制在每个区域的树冠覆盖范围内，这在数量、位置和孤立树的大小。对旱地单株树木量化的关注有限，导致对冠层覆盖范围的误解，并导致与冠层覆盖定义相关的混淆（即“覆盖”计算中包含的木本植物特征）。旨在评估全球树木覆盖率的产品在量化旱地的树木覆盖率方面设计不佳，这导致普遍认为撒哈拉或萨赫勒等旱地地区基本上没有树木。
研究发现
A. 非森林林木识别
Here we present a wall-to-wall identification of non-forest trees (defined as woody plants with a crown size over 3 m²) in the West African Sahara, Sahel and sub-humid zone, covering a rainfall gradient from hyper-arid (rainfall of 0–150 mm yr−1), arid (rainfall of 150–300 mm yr−1), semi-arid (rainfall of 300–600 mm yr−1) to sub-humid (rainfall of 600–1,000 mm yr−1) areas. We split the crown sizes into shrubs (3–15 m²), small trees (15–50 m²) and large trees (50–200 m²), as well as very large trees and clumped canopies that form thickets or forests (over 200 m²).
在这里，我们提出了非森林树木的全面识别（定义为树冠大小超过 3 m2 的木本植物）在西非撒哈拉、萨赫勒和半湿润地区，覆盖从超干旱（降雨量为 0-150 mm yr−1）到干旱（降雨量为150–300 mm yr−1)、半干旱（降雨量为 300–600 mm yr−1）到半湿润（降雨量为 600–1,000 mm yr−1）地区。我们将树冠大小分为灌木（3-15 平方米）、小树（15-50 平方米）和大树（50-200 平方米），以及非常大的树木和形成灌木丛或森林的丛生树冠（超过 200 平方米）。
Figure 1. Mapping trees using deep learning. a, Forests in a previously published global tree-cover map are defined as more than 25% canopy closure of trees taller than 5 m. This definition does not apply in most dryland areas, as in these regions trees grow mostly as isolated plants. This study mapped all trees (>3 m2 crown size)in the red rectangle using deep learning applied to submetre-resolution satellite imagery. b, The density of trees per hectare along the rainfall gradient (0–1,000 mm yr-1), derived from 1,837,565,501 trees. The cyan lines are the 150-, 300-, 600- and 1,000-mm-per-yr rainfall isohyets (mean 1982–2017), increasing from north to south.
图 1. 深度学习绘制林木分布：(a) 树冠超过5m的树木， (b) 预测灌木 ，根据降雨梯度（ 150 毫米、300 毫米、600 毫米和 1,000 毫米）预测每公顷树木数量
Figure 2. Predicting tree crowns. This set of 256 × 256-pixel plots from the independent test dataset shows the capabilities of the convolutional neural network model to predict trees (right column) from panchromatic images (left column) and NDVI (central column) at 0.5-m resolution.
图 2. 预测树冠。这组来自独立测试数据集的 256 × 256 像素图显示了卷积神经网络模型从全色图像（左列）和 NDVI（中心列）以 0.5 米分辨率预测树木（右列）
B. 树木密度和覆盖率随着降雨梯度下降而降低
The results show that tree density and coverage develop along the rainfall gradient (Fig. 3) from very sparse in the hyper-arid Sahara Desert in the north (0.7 ± 2.6 (mean ± s.d.) trees per hectare, and 0.1 ± 4% canopy cover), through scattered trees in arid (on average 9.9 ± 13.2 trees per hectare, and 1.6 ± 2.9% cover) and semi-arid (on average 30.1 ± 29.0 trees per hectare, and 5.6 ± 5.9% cover) lands, into denser patterns in the sub-humid south (on average 47 ± 30 trees per hectare, and 13.3 ± 9.4% cover).
结果表明，树木密度和覆盖率随着降雨梯度（图 3）从北部极度干旱的撒哈拉沙漠非常稀疏（每公顷 0.7 ± 2.6（平均 ± 标准差）树木和 0.1 ± 4% 的树冠覆盖率），穿过干旱地区的零星树木（平均每公顷 9.9 ± 13.2 棵树和 1.6 ± 2.9% 的覆盖率）和半干旱（平均每公顷 30.1 ± 29.0 棵树和 5.6 ± 5.9% 的覆盖率）土地，在半湿润的南部（平均 47每公顷 ± 30 棵树，覆盖率为 13.3 ± 9.4%）。
Figure 3. Cover and density of individual trees (a)The stacked contribution of the crown size classes to canopy cover is shown along the rainfall gradient (1-mm steps). (b)Canopy cover in relation to annual rainfall (10-mm steps). (c)for sandy soils (>70% sand content) and up to 800-mm rainfall (owing to low coverage of sandy areas beyond this threshold). (d)Tree density along the rainfall gradient (1-mm steps), stacked into crown size classes. (e)Tree density along the rainfall gradient (10-mm steps). (f) ut for sandy soils.
图 3. 单棵树的覆盖和密度 (a) 树冠大小等级对冠层覆盖的叠加贡献沿降雨梯度; (b)与年降雨量相关的冠层覆盖（10 毫米步长）;(c)对于沙质土壤（>70% 的沙子含量）和高达 800 毫米的降雨量(由于超出此阈值的沙地覆盖率较低); (d)沿降雨梯度的树木密度，堆叠成树冠大小等级;(e)沿降雨梯度的树木密度（10 毫米步长）;(f)适用于沙质土壤。
结论
Stabilizing the crown size of each tree more than 3 m² in size over a land area that spans 1.3 million km² in the West African Sahara, Sahel and sub-humid zone, using submetre-resolution satellite imagery and deep learning. We detected over 1.8 billion individual trees (13.4 trees per hectare), with a median crown size of 12 m², along a rainfall gradient from 0 to 1,000 mm per year. The canopy cover increases from 0.1% (0.7 trees per hectare) in hyper-arid areas, through 1.6% (9.9 trees per hectare) in arid and 5.6% (30.1 trees per hectare) in semi-arid zones, to 13.3% (47 trees per hectare) in sub-humid areas. Although the overall canopy cover is low,the relatively high density of isolated trees challenges prevailing narratives about dryland desertification, and even the desert shows a surprisingly high tree density. Our assessment suggests a way to monitor trees outside of forests globally, and to explore their role in mitigating degradation, climate change and poverty.
实现使用亚米分辨率卫星图像和深度学习，在西非撒哈拉、萨赫勒和半湿润地区 130 万平方公里的土地上绘制了每棵树的树冠大小超过 3 平方米的地图。检测到超过 18 亿棵树（每公顷 13.4 棵树），平均树冠大小为 12 平方米，每年降雨梯度为 0 至 1,000 毫米。树冠从超干旱地区的 0.1%（每公顷 0.7 棵树）增加到 1.6%（每公顷9.9 棵树）在干旱地区和 5.6%（每公顷 30.1 棵树）在半干旱地区达到 13.3%（每公顷 47 棵树）在亚湿润地区。虽然整体的树冠覆盖率很低，相对高密度的孤立树木挑战了关于旱地荒漠化，甚至沙漠也显示出惊人的高树木密度。我们的评估提出了一种监测全球森林以外树木的方法，并探索它们在缓解退化、气候变化和贫困方面的作用。
References:
Brandt, M., Tucker, C.J., Kariryaa, A. et al. An unexpectedly large count of trees in the West African Sahara and Sahel.
Nature 587, 78–82 (2020). https://doi.org/10.1038/s41586-020-2824-5
参会方式：B站“GeoInsider”直播间平台，直播间可现场留言提问，进入方式如下二选一：
(a) B站直播间链接：https://live.bilibili.com/22243531
(b) B站直播间二维码:
如有侵权，请联系本站删除！
