Quantifying climate change impacts and predictability to inform climate risk assessment and adaptation

Frances Davenport

Colorado State University

Tuesday, Nov 05, 2024, 2:00 pm MT
DSRC Room 2A305

JOIN THE WEBINAR:

Google Meet
Or by phone: (US)+1 4702852247
Access Code ‪718 983 414‬#
More phone numbers: https://tel.meet.com/tfh-jikn-jxu

Abstract

Climate variability and long-term climate change create significant risks and challenges for a range of human and natural systems. However, explicitly quantifying these risks requires understanding how climate processes affect different socioeconomic or environmental systems. Additionally, effective adaptation requires understanding how the climate may evolve across a range of timescales.

In this talk, I will give examples of research to address each of these challenges. First, I will show examples using empirical econometric modeling to quantify damages from climate change, specifically in the context of flood damages and agricultural losses. Second, I will show recent work to explore multi-year predictability in the climate system using neural networks and large ensemble climate model simulations. We find that neural networks can identify “windows of opportunity” where future sea surface temperature (SST) anomalies can be predicted with more certainty. Additionally, neural networks trained on climate models also make skillful SST predictions in reconstructed observations, although the skill varies by region and depending on which climate model the network was trained. Overall, this highlights a potential new avenue to develop actionable climate predictions.

Bio: Frances Davenport (she/her) is an Assistant Professor of Civil and Environmental Engineering at Colorado State University. She is interested in the dynamics of global climate change, extreme climate events, and the hydrologic cycle, as well as understanding how climate change impacts human and natural systems. Her work uses multiple methods, including observational data from in-situ monitoring networks and remote sensing; global climate models; and computational analysis tools, including machine learning.


Seminar Contact: psl.seminars@noaa.gov