Gro’s New Climate Ensemble Model

The Gro Climate Ensemble Model offers a significant improvement over a single climate model in forecast accuracy and precision

Climate change-related disasters, like floods, wildfires, and hurricanes, come with high human and economic costs, and until addressed, the scope and the materiality of these physical climate risks will grow. 

Because reliable climate forecasting is integral to physical climate risk identification and management, Gro Intelligence’s in-house team of climate and data scientists have spent the past year building our newest climate projection model, the Gro Climate Ensemble Model.

Starting in the beginning of April, the Gro Climate Ensemble will power the Gro Climate Risk Navigators and the Gro API.

This model, by combining models from multiple independent research efforts, offers a significant improvement over a single climate model in forecast accuracy and precision, allowing greater confidence at a finer resolution.

The Gro Climate Ensemble uses raw climate data provided by a subset of labs that participated in the Intergovernmental Panel on Climate Change’s (IPCC) latest global climate change report to produce climate forecasts that reflect the collective knowledge of the world’s top climate research labs. 

Our new model also conveys to users how confident the global climate research community is about each climate projection out to 2100 and under various global warming scenarios. 

To conduct the kind of Task Force on Climate-Related Financial Disclosures (TCFD) -aligned physical climate risk disclosures that regulators are increasingly asking for, public companies need credible physical climate risk data. 

Only when a company, asset manager, or government has a clear sense of the likelihood of a climate hazard can it assign an expected dollar value to a physical climate risk exposure and only then can it justify expenditures to address that physical climate risk exposure. Companies,  asset managers, and governments can also use this information to identify investment opportunities that are resilient to these physical climate risks.

Gro’s Ensemble Model is an IPCC-based solution 

For its Sixth Assessment Report (AR6), the IPCC has been using the outputs of the more than 50 different climate models contained in the CMIP6 multi-model ensemble to estimate the amount of global warming associated with five different global warming scenarios. 

 To distill all of this CMIP6 data into clear, consensus-based climate change signals, our climate scientists built the Gro Climate Ensemble. Our Climate Ensemble averages the projections of nine of the more than 50 CMIP climate models - after considering the sensitivity of each CMIP6 participating lab’s models, the completeness and frequency of each research lab’s data (specifically, the availability of all environmental variables needed to model climate impacts), and the number of global warming scenarios that each lab simulated. 

A guiding principle while developing the Gro Climate Ensemble was that vetting the CMIP6 participating labs individually and then averaging the results of the labs that produced the most robust and complete data sets would tease out consensus-based climate change signals that contain less statistical noise. 

After identifying a subset of climate models, our climate and data scientists then bias-adjusted the nine CMIP6 model predictions that they selected using historical reanalysis data sets and aggregated Gro Climate Ensemble’s climate projections to district (county), state/province, and national levels globally. 

The Climate Ensemble’s geographical language aligns with the geographical language used across the Gro platform, enhancing geospatial useability and interoperability across all industries. From April, the Gro Climate Ensemble will power the Gro Climate Risk Navigator and the Gro API, beginning with temperature and precipitation projections for the five global warming scenarios used in AR6.  

In the coming months, we will be upgrading our existing Gro Climate Indicator Projection (GCIP) data series (e.g. Weather Variation, Extreme Heat, Heavy Precipitation) to run on the Gro Climate Ensemble. All future GCIPs (cyclones, drought, flood, sea level rise, etc.) will be built on our Ensemble as well.

Over the next few weeks, we look forward to telling you more about how we developed the Gro Climate Ensemble, and how it can be used in scenario analysis and stress testing as markets embed physical climate risk data into decision-making processes and prepare TCFD-aligned physical climate risk disclosures.