Correcting temperature patterns in Integrated Assessment Models: Implementing local air quality and land cover temperature effects in IMAGE
Summary
The Dutch Environmental Assessment Agency (PBL) uses an integrated assessment model (IAM), IMAGE, to address a set of global environmental issues and sustainability challenges. IMAGE is unique compared to other IAMs because of its high spatial resolution for land-based processes. Among others, IMAGE is used to develop a set of unique scenarios, the Shared Socio-economic pathways (SSPs), to facilitate the integrated analysis of future climate impacts, vulnerabilities, adaptation, and mitigation.
In order to model temperature change at high resolution, a two-step approach is used. First, global temperature change is modelled with a reduced complexity climate model, which is scaled down using Earth System Model (ESM) results in a second step. However, SSPs studied with IMAGE can significantly deviate from the current downscaling scenarios, particularly in terms of air quality (AQ) and land cover (LC). This implies that the down-scaled temperature patterns only account for global effects of aerosol and LU-change effects, while strong local effects of AQ and LC changes are expected.
This study focuses on improving the temperature patterns by implementing temperature corrections for local AQ and LC effects. For AQ, a source receptor atmospheric chemistry model is used to model the concentration fields for aerosol components. Future concentrations are subsequently scaled with a set of controlled experiments and its temperature changes. For LC, aggregated remote sensing data on local biophysical LC change effects is used.
When the correction models are compared to ESMs in validation experiments, the AQ corrections produce results which are more or less in line with ESM results. Comparisons with fixed SST ESM runs show better agreement, with correlation coefficients up to 0.88. For LC, correlation between the corrections is small at best for some ESMs and absent in most cases.
Applying these methods on the newly developed SSPs, local AQ corrections of 0.5 K and up to 1 K are calculated in South-West, South and East Asia. For LC, the correction magnitudes are smaller and locally varying and primarily found in South America Africa. Global average temperature is only marginally affected by the corrections. Applying the temperature corrections in the IMAGE land model leads to changes in the order of 1\% for crop yields and living biomass in China and India.