Coastal stratocumulus clouds (Sc) have unique modeling challenges due to their development over coastal land, one of them being the accurate representation of surface fluxes. Unlike marine Sc, where the ocean can store significant heat and release relatively constant surface heat fluxes over the day, there are strong diurnal variations of both sensible and latent heat fluxes over land. Moreover, land surface fluxes have a strong feedback with cloud cover. Many modeling efforts have been directed to improve the representation of surface fluxes through developing more accurate land surface models with increasing complexity. Regarding the boundary layer turbulence, for marine Sc, greater sensible fluxes are known to intensify updrafts and increase entrainment, while greater latent heat fluxes have been linked to decoupling. An example of surface flux variations for Marine Sc is the transition of Sc to shallow Cumulus along the trade winds, which occurs over a number of days. For coastal land, changes of surface fluxes occur in a much shorter timescale (hours), and the sensitivity of their dynamical response has not been explored. In this work, we study the response of coastal Sc to controlled variations of surface fluxes using Large Eddy Simulations. Representative scenarios of diurnal profiles are generated using 12 years of surface flux measurements for cloudy days over southern California, and then simulated under several configurations that describe sudden and gradual changes of surface fluxes with varied timing and magnitude. Sudden changes result in increased cloud thinning and earlier dissipation times, although the timing of the sudden increase is also important, in relation to the original dissipation time. The response time to sudden changes of surface fluxes is evident in the evolution of maximum vertical velocity and vertically integrated Turbulent Kinetic Energy, with timescales of 1 and 2 eddy turnover times, respectively.

Recent insights into the spatial organization of atmospheric convection have emphasized the importance of its correct representation in Earth System Models (ESM). This study explores new opportunities created when combining a thermal population model on a horizontal microgrid with a decentralized vertical transport model. To this purpose the recently proposed BiOMi population model (Binomials on Microgrids) is used. BiOMi mimicks a population of independent but interacting convective thermals, with their birth, movement and life cycle described as Bernoulli processes. Simple rules of interaction are introduced to reflect observed physical behavior in single cumulus clouds, such as pulsating growth and environmental deformation. Under these rules, thermals can congregate and form longer-lived coherent clusters or chains that resemble cumulus clouds. The formation and evolution of these clusters is a form of self-organization that retains convective memory. Through an online clustering method the microgrid is coupled to a spectral EDMF convection scheme, providing the cluster size distribution it needs as input. This way, the inherently 3D structure of organized convection can in principle be captured in reduced but efficient form. The system is fully decentralized in that central top-down bulk closures are avoided. The main science objective of this study is to provide proof of concept of decentralized frameworks of this kind. To this purpose the BiOMi-EDMF scheme as implemented in the DALES circulation model is tested for various LASSO cases of shallow convection at the ARM SGP site. We find that the scheme achieves stable and realistic diurnal quasi-equilibria (as shown in the figure), and that the associated self-organizing patterns on the microgrid are realistic. Impacts of spatial organization and convective memory on the parameterized transport will be investigated.

The convective nature of Stratocumulus topped boundary layers (STBL) involves the motion of updrafts and downdrafts, driven by surface fluxes and radiative cooling, respectively. The balance between shear and buoyant forcings at the surface can determine the organization of updrafts between cellular and roll structures. We investigate the effect of varying shear at the surface and top of the STBL using LES simulations, taking DYCOMS II RF01 as a base case. We focus on spatial identification of the following features: coherent updrafts, downdrafts, and wet updrafts, and observe how they are affected by varying shear. Stronger surface shear organizes the updrafts in rolls, causes less well-mixed thermodynamic profiles, and decreases cloud fraction and LWP. Stronger top shear also decreases cloud fraction and LWP more than surface shear, by thinning the cloud from the top. Features with stronger top than surface shear are associated with a net downward momentum transport and show early signs of decoupling. Classifying updrafts and downdrafts based on their vertical span and horizontal size confirms the dominance of large objects spanning the whole STBL. Large objects occupy 14% of the volume in the STBL while smaller ones occupy less than 1%. For updraft and downdraft fluxes these large objects explain 33% of the vertical velocity variance and 53% of the buoyancy flux, on average. Stronger top shear also weakens the contribution of downdrafts to the turbulent fluxes and tilts the otherwise vertical development of updrafts.

A clustering method is applied to high resolution simulations of shallow continental convection to investigate the size dependence of coherent structures in the convective boundary layer. The study analyses the geometry of the clusters, along with their profiles of vertical velocity and total water. The main science goal is to assess various assumptions often used in spectral mass-flux convection schemes. Novel aspects of the study methodology include i) a newly developed clustering algorithm, and ii) an unprecedentedly large number of simulations being analysed. In total 26 days of LASSO simulations at the ARM-SGP site are analyzed, yielding roughly one million individual clusters. Plume-like surface-rooted coherent convective clusters are found to be omnipresent, the depth of which is strongly dependent on cluster size. The largest clusters carry vertical structures that are roughly consistent with the classic buoyancy-driven rising plume model, while smaller clusters feature considerable variation in top height. The cluster area is found to strongly vary with height and size, with small clusters losing mass and large clusters gaining mass below cloud base. Similar size dependence is detected in kinematic and thermodynamic properties, being strongest above cloud base but much weaker below. Finally the efficiency of the top-hat approach in flux parameterization is investigated, found to be 80-85 \% including a weak but well-defined dependence on cluster size. Implications of the results for spectral convection scheme development are briefly discussed.

Coastal Stratocumulus clouds (Sc) can greatly impact the available solar radiation in areas like southern California. These clouds usually dissipate during the morning, as surface fluxes increase in the day. The dissipation process begins by the thinning of the cloud layer, that then becomes a broken cloud field until the clouds completely dissipate. A better understanding of the process could improve predictions of dissipation time and solar variability for this type of clouds. We investigate the evolution of a broken cloud field using LES simulations of coastal Sc clouds over Southern California (Ghonima et al, JAS 2016). We identify and characterize coherent updrafts, downdrafts, and cloud elements using flow field variables (vertical velocity and liquid water mixing ratio). The cloud layer is observed to become a broken cloud field around hour 8 of the simulation, after which the thinning tendency strongly diminishes and the variability of cloud fraction, cloud thickness, and liquid water path suddenly increases. The few cloud elements remaining at noon are found to be located close to cloud base, unlike at night, when they are distributed throughout the whole cloud layer. The largest updrafts and downdrafts contribute to a large portion of the turbulent fluxes along the day (on average, 40% of vertical velocity variance and 48% of the buoyancy flux), and their relative contribution changes during the day as updrafts become stronger and downdrafts weaken around noon.

Vertical cloud overlap, the ratio of cloud fraction by area and by volume, for cumulus clouds are studied using large-eddy simulations (LES) due to the inefficient, wide-range values of cloud overlap. We can obtain information about the cloud cover of a cloud field by inspecting the individual clouds in that cloud field.We start with the maximum-random assumption and adjust this assumption for individual clouds.From this there is an under-prediction which leads to the conclusion that something can be added.We extend this by considering physical factors of cloud overlap: area variability, vertical wind shear, and turbulence. We use numerical schemes to calculate the effect of each contributor based on cloud height. We obtain great accuracy for our model of cloud overlap. Since there are multiple factors of cloud overlap, we look at the percentage of how much each contributes for a given binned cloud height. Furthermore, we get acceptable agreement for the calculated and actual total cloud cover. As such, we show that no other major contributors for cloud overlap and cloud cover exist. We end with an empirical model to describe the numerical schemes mentioned previously.

Multi-plume convection schemes based on discretized size densities can be easily made scale-aware by size filtering the parametrized plumes to fit the model resolution. Combined with their capability to smoothly transition from shallow to deep convection while being both numerically robust and computationally viable, these bin macrophysics models are well suited to the needs of high resolution NWP and regional climate modeling. One such approach is the ED(MF)n scheme which has previously been evaluated in a single column model against maritime cumulus LES. In this presentation we will confront the ED(MF)n scheme with long-term Doppler lidar observations of shallow cumulus days from the ARM-SGP site in Oklahoma. To do so the ED(MF)n scheme is implemented and tested in a non-hydrostatic model on a microgrid of 8x8 or 16x16 gridpoints. By using a ~10 km resolution this system behaves as a collection of weakly interacting single column models. We will compare the modeled relationship of vertical velocity to plume size against the relationship found in both the Doppler lidar observations and the LASSO LES run routinely at the ARM-SGP site. Next we examine which effect the assumed cloud-size distribution and entrainment as a function of plume size have on the resulting vertical transport.