| description abstract | We assess the vertical distribution of radiative heating rates (RHRs) in climate models using a multimodel experiment and A-Train satellite observations, for the first time. As RHRs rely on the representation of cloud amount and properties, we first compare the modeled vertical distribution of clouds directly against lidar?radar combined cloud observations (i.e., without simulators). On a near-global scale (50°S?50°N), two systematic differences arise: an excess of high-level clouds around 200 hPa in the tropics, and a general lack of mid- and low-level clouds compared to the observations. Then, using RHR profiles calculated with constraints from A-Train and reanalysis data, along with their associated maximum uncertainty estimates, we show that the excess clouds and ice water content in the upper troposphere result in excess infrared heating in the vicinity of and below the clouds as well as a lack of solar heating below the clouds. In the lower troposphere, the smaller cloud amount and the underestimation of cloud-top height is coincident with a shift of the infrared cooling to lower levels, substantially reducing the greenhouse effect, which is slightly compensated by an erroneous excess absorption of solar radiation. Clear-sky RHR differences between the observations and the models mitigate cloudy RHR biases in the low levels while they enhance them in the high levels. Finally, our results indicate that a better agreement between observed and modeled cloud profiles could substantially improve the RHR profiles. However, more work is needed to precisely quantify modeled cloud errors and their subsequent effect on RHRs. | |
