Error Measurement and Propagation: Understanding the Reliability of the Wildfire Resilience Index

Abstract:
<p>Background:&nbsp;As wildfires continue to grow in scale and intensity, larger and larger populations are at-risk of experiencing a wildfire. Population demographics play an important role in the recovery and preparedness of communities in wildfire prone areas. As such, it can be useful to have social vulnerability indices to understand the potential risk for communities and make appropriate decisions about resources and emergency management. The National Center for Ecological Analysis and Synthesis has recently published a “wildfire resilience index” that aims to capture local wildfire risk, preparedness, and social vulnerability. Purpose: we investigate the sensitivity of the WRI to the error of its component parts (American Community Survey, etc.) to better understand the range of possible outcomes and value as an interpretative / decision-making tool. Methods: we use statistical resampling methods to appropriately propagate errors from underlying data structures through to the final index and assess. Conclusion: While this work is preliminary, we suspect that the WRI will be highly sensitive to changes in the underlying data assumptions, and less reliable for decision making. Further, inattention to characterization of sampling variability in components of the WRI will lead to false statements of precision about the location and level of vulnerable populations. This paves this way for the development of new, targeted indices for wildfire resilience and social vulnerability.&nbsp;</p>

Keywords: Wildfire, social vulnerability index, demography, uncertainty, resampling

Authors: