The Cloud System Evolution in the Trades (CSET) study was designed to describe and explain the evolution of the boundary layer aerosol, cloud, and thermodynamic structures along trajectories within the North Pacific trade winds. The study centered on seven round trips of the National Science Foundation–National Center for Atmospheric Research (NSF–NCAR) Gulfstream V (GV) between Sacramento, California, and Kona, Hawaii, between 7 July and 9 August 2015. The CSET observing strategy was to sample aerosol, cloud, and boundary layer properties upwind from the transition zone over the North Pacific and to resample these areas two days later. Global Forecast System forecast trajectories were used to plan the outbound flight to Hawaii with updated forecast trajectories setting the return flight plan two days later. Two key elements of the CSET observing system were the newly developed High-Performance Instrumented Airborne Platform for Environmental Research (HIAPER) Cloud Radar (HCR) and the high-spectral-resolution lidar (HSRL). Together they provided unprecedented characterizations of aerosol, cloud, and precipitation structures that were combined with in situ measurements of aerosol, cloud, precipitation, and turbulence properties. The cloud systems sampled included solid stratocumulus infused with smoke from Canadian wildfires, mesoscale cloud–precipitation complexes, and patches of shallow cumuli in very clean environments. Ultraclean layers observed frequently near the top of the boundary layer were often associated with shallow, optically thin, layered veil clouds. The extensive aerosol, cloud, drizzle, and boundary layer sampling made over open areas of the northeast Pacific along 2-day trajectories during CSET will be an invaluable resource for modeling studies of boundary layer cloud system evolution and its governing physical processes.

Despite steady improvements in the skill of numerical weather and climate models over the last decades, a longstanding issue is the development of biases after initialization. These biases (systematic forecast errors) cause degradation of performance for both medium range weather forecasting and subseasonal to decadal climate predictions. They arise from issues like limited resolution, inaccurate physical parameterizations, and imperfect initial conditions. Typically, postprocessing steps are developed to handle these biases such as model output statistics for weather forecasting (Glahn & Lowry, 1972) or ensemble bias correction for seasonal prediction (Arribas et al., 2011;Stockdale et al., 1988). In this study, we propose an online bias correction method using machine learning (ML). We apply a corrective tendency to the prognostic state of the atmospheric model at each time step in order to reduce atmospheric model error growth. The necessary corrective tendencies are estimated from a hindcast simulation which is linearly nudged towards an observational analysis. An ML model is trained to predict the nudging tendencies using only the state of the model as inputs. This ML model can then be used in a forecast to keep the model evolution on a more realistic manifold.

Global atmospheric “storm‐resolving” models with horizontal grid spacing of less than 5 km resolve deep cumulus convection and flow in complex terrain. They promise to be reference models that could be used to improve computationally affordable coarse‐grid global climate models across a range of climates, reducing uncertainties in regional precipitation and temperature trends. Here, machine learning of nudging tendencies as functions of column state is used to correct the physical parameterization tendencies of temperature, humidity, and optionally winds, in a real‐geography coarse‐grid model (FV3GFS with a 200 km grid) to be closer to those of a 40‐day reference simulation using X‐SHiELD, a modified version of FV3GFS with a 3 km grid. Both simulations specify the same historical sea‐surface temperature fields. This methodology builds on a prior study using a global observational analysis as the reference. The coarse‐grid model without machine learning corrections has too few clouds, causing too much daytime heating of land surfaces that creates excessive surface latent heat flux and rainfall. This bias is avoided by learning downwelling radiative flux from the fine‐grid model. The best configuration uses learned nudging tendencies for temperature and humidity but not winds. Neural nets slightly outperform random forests. Forecasts of 850 hPa temperature gain 18 hr of skill at 3–7 days leads and time‐mean precipitation patterns are improved 30% by applying the ML correction. Adding machine‐learned wind tendencies improves 500 hPa height skill for the first five days of forecasts but degrades time‐mean upper tropospheric temperature and zonal wind patterns thereafter.

During the Marine ARM GPCI Investigation of Clouds (MAGIC) in October 2011 to September 2012, a container ship making periodic cruises between Los Angeles, CA, and Honolulu, HI, was instrumented with surface meteorological, aerosol and radiation instruments, a cloud radar and ceilometer, and radiosondes. Here large‐eddy simulation (LES) is performed in a ship‐following frame of reference for 13 four day transects from the MAGIC field campaign. The goal is to assess if LES can skillfully simulate the broad range of observed cloud characteristics and boundary layer structure across the subtropical stratocumulus to cumulus transition region sampled during different seasons and meteorological conditions. Results from Leg 15A, which sampled a particularly well‐defined stratocumulus to cumulus transition, demonstrate the approach. The LES reproduces the observed timing of decoupling and transition from stratocumulus to cumulus and matches the observed evolution of boundary layer structure, cloud fraction, liquid water path, and precipitation statistics remarkably well. Considering the simulations of all 13 cruises, the LES skillfully simulates the mean diurnal variation of key measured quantities, including liquid water path (LWP), cloud fraction, measures of decoupling, and cloud radar‐derived precipitation. The daily mean quantities are well represented, and daily mean LWP and cloud fraction show the expected correlation with estimated inversion strength. There is a −0.6 K low bias in LES near‐surface air temperature that results in a high bias of 5.6 W m−2 in sensible heat flux (SHF). Overall, these results build confidence in the ability of LES to represent the northeast Pacific stratocumulus to trade cumulus transition region.