Precipitation Characteristics in the Community Atmosphere Model and Their Dependence on Model Physics and Resolution

Abstract

Precipitation amount (A), frequency (F), intensity (I), and duration (D) are important properties of precipitation, but their estimates are sensitive to data resolution. This study investigates this resolution dependence, and the influences of different model physics, by analyzing simulations by the Community Atmospheric Model (CAM) version 4 (CAM4) and version 5 (CAM5) with varying grid sizes from~0.25 to 2.0°. Results show that both CAM4 and CAM5 greatly overestimate F and D but underestimate I at all resolutions, despite realistic A. These biases partly result from too much parameterized (convective) precipitation with high F and D but low I. Different cloud microphysics schemes contribute to the precipitation differences between CAM4 and CAM5. The A, F, I, and D of convective and nonconvective precipitation react differently to grid-size decreases, leading to the large decreases in F and D but increases in the I for total precipitation as model resolution increases. This resolution dependence results from the increased probability of precipitation over a larger area (area aggregation effect, which is smaller than in observations) and the varying performance of model physics under changing resolution (model adjustment effect), which roughly enhances the aggregation-induced dependence. Finer grid sizes not only increase resolved precipitation, which has higher intensity and thus improves overall precipitation intensity in CAM, but also reduce the area aggregation effect. Thus, the long-standing drizzling problem in climate models may be mitigated by increasing model resolution and modifying model physics to suppress parameterized convective precipitation and enhance resolved nonconvective precipitation.

The Effect of Modeling Strategies on Assessments of Differential Warming Impacts of 0.5°C

“…There is growing evidence of added value of high resolution modeling in simulating the global hydrological cycle as well as precipitation extremes (Prein et al, 2016;Roberts et al, 2018;Van Der Wiel et al, 2016;Vannière et al, 2018;Wu et al, 2019). For extreme precipitation, specifically, heavier extreme precipitation is permitted overall across the globe in higher-resolution simulations (Chen & Dai, 2019;O'Brien et al, 2016;Wehner et al, 2014). This is linked to resolution-dependence of dynamics with increasingly extreme updrafts in finer grids (O'Brien et al, 2016;Rauscher et al, 2016;Yang et al, 2014).…”

A physically based precipitation separation algorithm for convection‐permitting models over complex topography

“…In contrast, the algorithm attributes the most intense precipitation to convective events out on the tails of the distribution. Admittedly, this could partly be explained by the coarse resolution of the dataset, since precipitation can be dependent on model resolution (Chen and Dai, ). However, this also points to a limitation of a classification based solely upon the parametrization scheme of a given model.…”

“…It was found that the model resolution could strongly affect the simulated vertical velocity, implying that the thresholds presented above are likely dependent on horizontal resolution within convection-permitting scales (e.g., O'Brien et al, 2013, Rauscher et al, 2016, Chen et al, 2019. The thresholds presented here were carefully selected to properly separate precipitation in these 3 km resolution simulations, but they would probably need to be increased for datasets at 1 or 2 km resolution.…”

A Model‐Based Investigation of Terrestrial Plant Carbon Uptake Response to Four Radiation Modification Approaches

“…Chen and Dai (2019) summarized that substantial differences exist between the CAM4 and CAM5 physics, including in the boundary layer turbulence scheme, shallow convection, cloud microphysics parameterizations, and radiation scheme. Such differences have been found to have impacts on precipitation and cloud cover responses to external forcing (Chen & Dai, 2019; Wall & Hartmann, 2015), and are expected to affect the surface climate condition and plant carbon uptake in response to various radiation modification approaches. The latest versions of atmosphere and land components used in CESM are CAM6 and CLM5 (e.g., Lawrence et al, 2019).…”