Monitoring Drought through the Lens of Landsat: Drying of Rivers during the California Droughts
Monitoring Drought through the Lens of Landsat: Drying of Rivers during the California Droughts
复制标题
DOI：
10.3390/rs13173423
复制
发表时间：
2021-08
期刊：
影响因子：
--
通讯作者：
Shang Gao;Zhi Li;Mengye Chen;D. Allen;T. Neeson;Yang Hong
中科院分区：
文献类型：
--
作者：
Shang Gao;Zhi Li;Mengye Chen;D. Allen;T. Neeson;Yang Hong
严重干旱期间的缺水对河流的水文和生态产生了深远的影响。然而，在干旱期间的河流表面范围的干燥动力学仍然在很大程度上研究不足。卫星遥感为现场观测提供了一种替代办法，从而能够以高空间分辨率对河流进行调查和分析。本研究调查了加州河流范围的季节性干旱动态，近几十年来严重干旱发生得越来越频繁。我们的方法结合使用基于陆地卫星的全球地表水（GSW）和全球河流堤岸宽度数据库联合收割机。作为一个间接的比较，我们研究了2071年河流河段的每月分数河的范围（FrcSA），并在同一地点的美国地质调查局测量流量的相关性。我们将2012-2015年的极端干旱置于数十年河流范围历史的更广泛背景下，并说明干旱期间和干旱后时期之间的非凡变化。除了河流范围动态，我们进行统计分析，将FrcSA与从国家陆地数据同化系统（NLDAS）模型模拟获得的水文气候变量联系起来。结果表明，Landsat提供了一致的观测超过90%的河流面积从3月至10月，是适合监测季节性河流干涸在加州。除干旱和山区外，FrcSA与径流量的相关性良好（>0.5）。在2012-2015年干旱期间，332条河流经历了Landsat历史34年来最低的年平均FrcSA。在月尺度上，FrcSA是更好地与土壤水分在更潮湿的地区。在年尺度上，夏季平均FrcSA是越来越敏感的冬季降水在干燥的气候，和弹性也降低了地下水位较深。总体而言，我们的研究表明，在干旱地区的河流表面的范围内，地形复杂的陆地卫星的可探测性。在未来气候更长、更频繁的干旱中，蓄水不足的集水区的河流范围可能会下降更高的百分比。
Water scarcity during severe droughts has profound hydrological and ecological impacts on rivers. However, the drying dynamics of river surface extent during droughts remains largely understudied. Satellite remote sensing enables surveys and analyses of rivers at fine spatial resolution by providing an alternative to in-situ observations. This study investigates the seasonal drying dynamics of river extent in California where severe droughts have been occurring more frequently in recent decades. Our methods combine the use of Landsat-based Global Surface Water (GSW) and global river bankful width databases. As an indirect comparison, we examine the monthly fractional river extent (FrcSA) in 2071 river reaches and its correlation with streamflow at co-located USGS gauges. We place the extreme 2012–2015 drought into a broader context of multi-decadal river extent history and illustrate the extraordinary change between during- and post-drought periods. In addition to river extent dynamics, we perform statistical analyses to relate FrcSA with the hydroclimatic variables obtained from the National Land Data Assimilation System (NLDAS) model simulation. Results show that Landsat provides consistent observation over 90% of area in rivers from March to October and is suitable for monitoring seasonal river drying in California. FrcSA reaches fair (>0.5) correlation with streamflow except for dry and mountainous areas. During the 2012–2015 drought, 332 river reaches experienced their lowest annual mean FrcSA in the 34 years of Landsat history. At a monthly scale, FrcSA is better correlated with soil water in more humid areas. At a yearly scale, summer mean FrcSA is increasingly sensitive to winter precipitation in a drier climate; and the elasticity is also reduced with deeper ground water table. Overall, our study demonstrates the detectability of Landsat on the river surface extent in an arid region with complex terrain. River extent in catchments of deficient water storage is likely subject to higher percent drop in a future climate with longer, more frequent droughts.