Precipitation, water level and descriptive statistics

Extreme weather events and the presence of mega-hydroelectric dams, when
combined, present an emerging threat to natural habitats in the Amazon
region. To understand the magnitude of these impacts, we used remote
sensing data to assess forest loss in areas affected by the extreme 2014
flood in the entire Madeira River basin, the location of two mega-dams. In
addition, forest plots (26 ha) were monitored between 2011 and 2015
(14,328 trees) in order to evaluate changes in tree mortality, aboveground
biomass (AGB), species composition and community structure around the
Jirau reservoir (distance between plots varies from 1 to 80 km). We showed
that the mega-dams were the main driver of tree mortality in Madeira basin
forests after the 2014 extreme flood. Forest loss in the areas surrounding
the reservoirs was 56 km² in Santo Antônio, 190 km² in Jirau (7.4-9.2% of
the forest cover before flooding), and 79.9% above that predicted in
environmental impact assessments. We also show that climatic anomalies,
albeit with much smaller impact than that created by the mega-dams,
resulted in forest loss along different Madeira sub-basins not affected by
dams (34-173 km²; 0.5-1.7%). The impact of flooding was greater in várzea
and transitional forests, resulting in high rates of tree mortality
(88-100%), AGB decrease (89-100%), and reduction of species richness
(78-100%). Conversely, campinarana forests were more flood-tolerant with a
slight decrease in species richness (6%) and similar AGB after flooding.
Taking together satellite and field measurements, we estimate that the
2014 flood event in the Madeira basin resulted in 8.81-12.47 ∙ 106 tons of
dead biomass. Environmental impact studies required for environmental
licensing of mega-dams by governmental agencies should consider the
increasing trend of climatic anomalies and the high vulnerability of
different habitats to minimize the serious impacts of dams on Amazonian
biodiversity and carbon stocks.

极端天气事件与巨型水力大坝（mega-hydroelectric dams）的叠加作用，正成为亚马逊地区自然栖息地面临的新兴威胁。为明晰此类影响的严重程度，我们借助遥感数据，对2014年极端洪水波及的整个马代拉河流域——两座巨型大坝的所在地——的森林丧失情况展开评估。此外，2011至2015年间，我们在吉拉乌水库（Jirau reservoir）周边布设了26公顷的森林监测样地，共监测14328株树木，以评估树木死亡率、地上生物量（aboveground biomass, AGB）、物种组成及群落结构的变化；各监测样地与水库的距离介于1至80千米之间。研究结果表明，在2014年极端洪水过后，巨型大坝是马代拉河流域森林树木死亡率的主要驱动因素。两座水库周边的森林丧失面积分别为：圣安东尼奥水库周边56平方千米，吉拉乌水库周边190平方千米，分别占洪水前森林覆盖率的7.4%至9.2%，且较环境影响评估的预测值高出79.9%。我们还发现，尽管气候异常对森林的影响远小于巨型大坝，但仍导致马代拉河流域未受大坝影响的子流域出现森林丧失，总面积达34至173平方千米，占对应区域洪水前森林覆盖率的0.5%至1.7%。洪水对瓦尔泽亚林（várzea）与过渡林的影响更为显著，带来了极高的树木死亡率（88%至100%）、地上生物量大幅下降（89%至100%）以及物种丰富度锐减（78%至100%）。与之相反，卡皮纳拉纳森林（campinarana）的耐涝性更强，仅出现小幅的物种丰富度下降（6%），洪水过后地上生物量与洪水前基本持平。综合卫星遥感与实地监测数据，我们估算2014年马代拉河流域的洪水事件共造成881万至1247万吨的死亡生物量。政府机构为巨型大坝开展环境许可审批所需的环境影响研究，应考虑到气候异常的加剧趋势，以及不同栖息地的高脆弱性，以最大限度降低大坝对亚马逊生物多样性与碳储量的严重影响。