Risk Modelling Part 2: Data

risk

series

data

Published

December 12, 2025

Modified

December 16, 2025

← Previous: Foundations

Next: Implementation →

Data Sources for Climate Risk

Building on the foundations from Part 1, this note covers the data landscape for climate risk modelling.

Observational Data

ERA5 - ECMWF’s reanalysis dataset
GHCN - Global Historical Climatology Network
Station data - Local weather observations

Modelled Data

CMIP6 - Climate model ensembles for projections
Downscaled projections - Regional climate scenarios
Catastrophe model outputs - Vendor-specific hazard layers

Data Quality Considerations

When working with climate data, always consider:

Spatial resolution - Is the data granular enough for your use case?
Temporal coverage - Do you have enough historical data?
Bias - Are there systematic errors in the observations?
Uncertainty - What are the confidence intervals?

Various frameworks provide approaches to validating spatial data quality.

Next Steps

With data in hand, we move to Part 3: Implementation where we build actual models.

Linked from

---
title: "Risk Modelling Part 2: Data"
date: 2025-12-12
categories: [risk, series, data]
---

::: {.series-nav}
::: {.series-nav-header}
Series: Introduction to Risk Modelling (Part 2 of 3)
:::
::: {.series-nav-links}
::: {.series-nav-prev}
[← Previous: Foundations](risk-series-1.html)
:::
::: {.series-nav-next}
[Next: Implementation →](risk-series-3.html)
:::
:::
:::

## Data Sources for Climate Risk

Building on the [foundations from Part 1](risk-series-1.html), this note covers the data landscape for climate risk modelling.

### Observational Data

- **ERA5** - ECMWF's reanalysis dataset
- **GHCN** - Global Historical Climatology Network
- **Station data** - Local weather observations

### Modelled Data

- **CMIP6** - Climate model ensembles for projections
- **Downscaled projections** - Regional climate scenarios
- **Catastrophe model outputs** - Vendor-specific hazard layers

## Data Quality Considerations

When working with climate data, always consider:

1. **Spatial resolution** - Is the data granular enough for your use case?
2. **Temporal coverage** - Do you have enough historical data?
3. **Bias** - Are there systematic errors in the observations?
4. **Uncertainty** - What are the confidence intervals?

Various frameworks provide approaches to validating spatial data quality.

## Next Steps

With data in hand, we move to [Part 3: Implementation](risk-series-3.html) where we build actual models.

::: {.backlinks}
#### Linked from
- [Risk Modelling Part 1: Foundations](risk-series-1.html)
- [Risk Modelling Part 3: Implementation](risk-series-3.html)
:::