Most investigations into the sensitivity of climate simulations to the horizontal resolution have identified some systematic improvements related to increases in horizontal resolution. The improvements can be due to a number of factors, including a more accurate representation of the topography, and a better simulation of the large-scale circulation. The expectation is that increasing spatial resolution will generally cause the simulation to improve, this is not always the case in practice. One issue is that the model is tuned to give a realistic climate at a specific resolution. A finer resolution model could give worse results unless it is retuned at that resolution. Another issue is that approximations made in the model can be violated if the resolution increases far enough.

Here we investigate the impact of increased resolution in the Community Atmospheric Model (CAM). We explore a range of horizontal resolutions from 2 degree to 0.25 degree. Twenty years of uncoupled simulations using observed sea surface temperatures (SSTs) and observed sea-ice are used to characterize the climate response to different resolutions.

Overall, the high-resolution simulation exhibits a mixed message. Some features are clearly improved e.g., the better resolution of topography leads to improved precipitation distributions in the Indian monsoon region. However, other features in the precipitation pattern are degraded such as the Philippine Sea monsoon extension in the 0.25 degree simulation. Other errors remain unchanged with varying resolution such as the double ITCZ and summertime US precipitation patterns.

In this paper, we document the climate response to the resolution change and we explore the reasons behind the changes.