EGU2020-1850, updated on 31 Dec 2020
https://doi.org/10.5194/egusphere-egu2020-1850
EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Spatial dependence of floods and droughts: learning from differences in regional and seasonal patterns

Manuela Irene Brunner1, Eric Gilleland1, Daniel Swain2, Andy Wood1, and Martyn Clark3
Manuela Irene Brunner et al.
  • 1National Center for Atmospheric Research, Research Applications Laboratory, Boulder, United States of America
  • 2University of California, Los Angeles, CA, United States of America
  • 3College of Arts and Science, University of Saskatchewan, Canmore, Canada

Regional flood and drought events often have more severe impacts than localized events in terms of damages and costs, the number of affected people, and habitat changes. Understanding which regions may be jointly affected by such extreme events can help us to derive reliable regional risk estimates, plan and manage resource flows, and develop suitable adaptation measures. However, the spatial dimension of droughts and floods is often neglected when deriving hazard estimates and we know little about the processes governing their spatial dependencies. Therefore, we investigate how and why spatial dependencies in droughts and floods vary seasonally and regionally over the United States. We aim to gain new insight into processes governing spatial dependencies of droughts and floods by contrasting their regional and seasonal patterns.

To map regions with a similar seasonal flood and drought behavior, respectively, we introduce a measure of connectedness, which quantifies the number of catchments with which a specific catchment co-experiences flood or drought events. We then summarize the spatial dependencies by identifying regions with a similar flood behavior and regions with a similar drought behavior. To do so, we use a hierarchical clustering procedure on the F-madogram, which is a measure of spatial dependence for extremes. We look at regional and seasonal differences in spatial dependence both for floods and droughts and subsequently compare the two phenomena.

We find that spatial dependence is over all seasons stronger for droughts than for floods. Both types of extremes, however, show regional and seasonal differences in spatial connectedness. Droughts show the strongest spatial dependence in fall. In contrast, the Rocky Mountains show the highest spatial dependence of droughts in winter because of snow accumulation. Very low spatial dependence is found in spring. The seasonal, spatial dependence patterns of floods are opposed to the one of droughts. Spatial flood dependence is highest in spring, especially in mountainous areas, high in winter at the Pacific coast and the Appalachian Mountains, and high in summer in the Rocky Mountains. In contrast, spatial connectedness is very weak in fall.
We conclude that spatial dependence patterns are stronger for droughts than floods because of the slower processes and longer durations associated with the phenomenon.   Furthermore, we conclude that both meteorological and land surface processes such as snowmelt and the availability of soil moisture shape the spatial dependence patterns of each extreme.

 

How to cite: Brunner, M. I., Gilleland, E., Swain, D., Wood, A., and Clark, M.: Spatial dependence of floods and droughts: learning from differences in regional and seasonal patterns, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1850, https://doi.org/10.5194/egusphere-egu2020-1850, 2019

Displays

Display file

Comments on the display

AC: Author Comment | CC: Community Comment | Report abuse

displays version 1 – uploaded on 30 Apr 2020
  • CC1: Comment on EGU2020-1850, Woonsup Choi, 06 May 2020

    The droughts were determined with the fixed threshold method. The spatial dependency would look quite different if variable thresholds were used. 

     

    • AC1: Reply to CC1, Manuela Irene Brunner, 07 May 2020

      Hi Woonsup,

      Yes, you are right, the spatial drought dependencies look different if one is using a variable threshold approach. It is all a matter of definition. I limited the presentation to the fixed threshold case.

      Manuela