World Weather Attribution
The World Weather Attribution Project is a collaborative project with Climate Central that aims to demonstrate the feasibility of near real-time attribution studies for extreme weather events around the world.
The goal of this ambitious project is to provide an immediate answer to the question: “to what extent did anthropogenic climate change play a role in an extreme weather event?” while that event is still unfolding and the world is waiting for an answer.
We will achieve this by using the weather@home experimental design, but with seasonal forecast inputs for the sea surface temperatures instead of observed ones, to simulate extreme weather events under current climate conditions one month ahead. This will enable us to compare them with the same events in the “world that might have been without climate change”. This means we will be able to say, almost in real-time, if the risk of an extreme weather event has been changed by climate change.
Since the first attribution study (Stott et al., 2004) was published, the science of probabilistic event attribution has seen formidable advances, not least through the work done at climateprediction.net. However, largely due to model deficiencies, our scientific understanding of some types of extreme events is limited. The types of events that we will be looking at in this project will thus focus on:
- Sea level rise contribution to storm surge (for hurricanes, typhoons, nor’easters, coastal storms)
- Extreme heat events
- Heavy rainfall events/flooding
By the end of this project we aim to have built the capacity to identify 1 in 50-year and 1 in 100-year events around the world and determine whether or not climate change increased the likelihood of that event.
Driving in heavy rain along the Grand Central Parkway, New York, by Steven Pisano
We are partnering with Climate Central, an independent organisation of leading scientists and journalists researching and reporting the facts about our changing climate and its impact on the public. Climate Central surveys and conducts scientific research on climate change and informs the public of key findings. Its scientists publish and its journalists report on climate science, energy, sea level rise, wildfires, drought, and related topics.
Climate Central currently provides weekly, localised climate content to more than 120 local weather forecasters across the United States. The primary goal of this project is to localise and contextualise extreme weather events as they relate to climate change.
Right now, it is very difficult for weather forecasters to quickly assess the frequency of a given hot day or heavy downpour as it is occurring. Was it a 1 in 10-year event or was it a 1 in 50-year event? Local weather forecasters need more information about this and this attribution initiative will address this need.
Real-time weather@home extreme event attribution
Our existing weather@home experiments run two sets of models: one representing the current climate as it was observed for recent extreme events in a particular region using observed sea surface temperatures (SSTs) to force the model, and the other to represent the same events in the “world that might have been” without human-induced climate change.
This project will change that basic setup by using seasonal forecast SSTs, rather than observed data, to simulate the “world that is”, while the “world that might have been” will be simulated as it is in other weather@home experiments, just using the seasonal forecast SSTs as the baseline.
The ultimate aim of this project is firstly to demonstrate the feasibility of using forecast instead of observed SSTs and still getting reliable attribution statements and secondly, to set things up so we can run models for all regions of the world for different types of extreme weather events, as they happen. This will involve the development of new regional models to cover regions we aren’t yet studying as well as sophisticated data management systems to deal with the very large data sets that will be produced by these experiments.