Work

Impacts of Model Resolution on Ocean Internal Variability in the Northern North Atlantic Analyzed within the Framework of AR(1)-Process

Year: 2026
Type: article
Abstract: Higher-resolution models explicitly resolve smaller-scale processes, thereby tending to enhance internal variability. While this perspective is well recognized and has been extensively investigated, a system-oriented perspective has received less attention and may offer complementary insight into the nature of internal variability. The system-oriented perspective refers to analyzing the climate system as a whole in terms of statistics. We explain the impacts of model grid resolution on the internal variability in the northern North Atlantic, focusing on two factors—memory and fluctuating forcing—within the framework of a first-order autoregressive (AR(1)) process, an exact form of the integral fluctuation–dissipation relation but associated with a single integration time interval. The ICON-ocean model is applied at grid resolutions of 20, 10, and 5 km, under identical daily climatological atmospheric forcing. Being captured by the AR(1)-framework, the total variance can be studied in terms of the two factors of the AR(1)-process: memory and fluctuating forcing. We find that the larger the variance of fluctuating forcing, the shorter the memory. Due to this general relationship, increasing horizontal resolution, allowing more variability to be resolved, leads to a strengthening of fluctuating forcing and a decrease in memory at the grid-point level. At both the grid-point level and across different spatial scales, increasing grid resolution has a stronger effect on the fluctuating forcing than on the memory, causing the total variance of internal variability to increase with increasing resolution.
Source: Tellus
Language: English
Subfield: Oceanography
Open Access status: diamond
OpenAlex ID: W7226030284
FWCI:
Cites: 18
Cited by:
Related to: 10