卡特里娜飓风对美国新奥尔良市西侧海岸线变化的影响

The impact of Hurricane Katrina on the coastline west of New Orleans, USA

卡特里娜飓风及其引发的风暴潮给路易斯安那州周围的海岸造成了灾难性的破坏.对海岸线进行连续观测有着重要意义,一方面可以用于定量评估飓风对海岸的破坏,另一方面有助于理解飓风对海岸的破坏过程.本文选取新奥尔良西侧一段约180 km的海岸线为研究区,综合使用美国陆地卫星时间序列和验潮站的水位数据得到飓风前后的海岸线变化率.提取不同时期的海岸线,然后进行速率估算和变化检测.结果表明,飓风登陆前后,海岸线位置和变化速率发生了较大的改变.卡特里娜飓风之前,研究区有39.42%的海岸线向海洋推进,60.58%的海岸线向陆地后退,海岸线平均变化速率为?2.53 m/a;而之后有26.73%的海岸线向海洋推进,73.27%的海岸线向陆地后退,海岸线平均变化速率为?3.58 m/a.卡特里娜飓风给海岸造成了很大破坏,86.57%的海岸在卡特里娜飓风中受到严重侵蚀,研究区海岸线向陆地平均移动14.91 m.计算结果表明遥感技术确实能够有效估算出飓风前后的海岸线变化.其结果可用于评估飓风对海岸的影响,为灾后重建提供基础信息,也可以用于研究飓风对海岸的破坏机理.本文显示了对地观测技术在灾害评估中发挥的巨大作用.

Forms of coastline are constantly modified by such driving forces as waves, sea level change, storms and winds. Among various driving forces, the effects of hurricanes can be the greatest and fastest. It is critical to continuously observe coastline changes in order to improve our understanding on the role of hurricane storms. One way is to make frequent field measurements along the coast before and after hurricane events. This is clearly labor-intensive at the regional scale. Remotely sensed data with sufficiently high acquisition frequency and spatial resolution can serve the purpose. Our question is if freely accessible Landsat data can be used to solve this problem. We used a 180-km long coast west of New Orleans that was affected by Hurricane Katrina as a case study. The storm surge induced by Hurricane Katrina caused well-known catastrophic damages in human lives and economic values. However, to what extent it has changed the coastline at the regional scale is an open question. We used Landsat time-series data（1984–2015） to estimate the rate of coastline change. First, the shortwave-infrared（1.55 mm） band of each Landsat image was selected to separate water and land. Water body detection was carried out using a simple threshold-based method. The optimal threshold was determined with a method based on the maximum inter-class variance criterion. Second, the boundary of the ocean water body was delineated as the coastline. Third, a linear regression model between water level and coastline position was established and an interpolation was conducted to remove the impact of clouds or shadows. To better analyze the impact of hurricanes on the coast, we removed the variation of water levels due to waves. Each average coastline positon was corrected to the long-term mean water level. The change rates were estimated during different observation periods. The results demonstrate that both the position and change rate of the coastline changed dramatically after the hurricane. Before Hurricane Katrina, 39% of the coastline advanced to ocean and 61% retreated to land, and the regional average change rate was ？2.53 m/a. After Hurricane Katrina, 27% of the coastline advanced to ocean and 73% retreated to land, and the regional average change rate was ？3.58 m/a. After Hurricane Katrina 87% of the coast suffered from serious erosion and the coastline shifted to land by 14.91 m during the hurricane. Further, we collected hurricane records in the study area from the National Hurricane Center and selected greater than level one hurricanes. We conducted an analysis of regime changes for coastline positions and found three regime turning points. We discovered that the regime turning points detected from Landsat time-series data matched well with the hurricane records. We conclude that there is a profound relationship between hurricane and coastline change. Landsat-level remote sensing data with 30 m spatial resolution can effectively estimate coastline changes caused by hurricanes in sediment dominated coastal areas. Coastline information derived from Landsat images could be used to assess impacts of hurricanes on the coast and to develop post-disaster reconstruction and restoration plans.