The paper provides a comprehensive review of the empirical values and assumptions used in the convection schemes of numerical models. The focus is on the values and assumptions used in the activation of convection (trigger), the transport and microphysics (commonly referred to as the cloud model), and the intensity of convection (closure). Such information can assist satellite missions focused on elucidating convective processes and the evaluation of model output uncertainties.

We proposed the leave-two-out method (i.e. one particular implementation of the nested cross-validation) to determine the optimal statistical crop model (using the validation dataset) and estimate its true generalization ability (using the testing dataset). This approach is applied to two examples (robusta coffee in Cu M'gar and grain maize in France). The results suggested that the simple models are more suitable in crop modelling where a limited number of samples is available.

Global ocean biogeochemical models are used within Earth system models which are used to predict future climate change. However, these are very computationally expensive to run and therefore are rarely routinely improved or calibrated to real oceanic observations. Here we apply a new, fast optimisation algorithm to one such model and show that it can calibrate the model much faster than previously managed, therefore encouraging further ocean biogeochemical model improvements.

In this paper, we develop an online data coupled assimilation system (NAQPMS-PDAF) with the Eulerian atmospheric chemistry-transport model. NAQPMS-PDAF allows efficient use of large computational resources. The application and performance of the system are investigated by assimilating 1 month of vertical aerosol observations. The results show that NAQPMS-PDAF can significantly improve the performance of aerosol vertical structure simulation and reduce the uncertainty to a large extent.

Mercury is a global pollutant that can be transported over long distance through the atmosphere. We develop a new online global model for atmospheric mercury. The model reproduces the observed global atmospheric mercury concentrations and deposition distributions by simulating the emissions, transport, and physicochemical processes of atmospheric mercury. And we find that the seasonal variations of atmospheric Hg are the result of multiple processes and have obvious regional characteristics.

The Arctic has large areas of small mounds that are caused by ice lifting up the soil. Snow blown by wind gathers in hollows next to these mounds, insulating them in winter. The hollows tend to be wetter, and thus the soil absorbs more heat in summer. The warm wet soil in the hollows decomposes, releasing methane. We have made a model of this, and we have tested how it behaves and whether it looks like sites in Scandinavia and Siberia. Sometimes we get more methane than a model without mounds.

This paper addresses numerical challenges in reasoning about geological models constrained by sensor data, especially models that describe the history of an area in terms of a sequence of events. Our method ensures that small changes in simulated geological features, such as the position of a boundary between two rock layers, do not result in unrealistically large changes to resulting sensor measurements, as occur presently using several popular modeling packages.

Progress in identifying complex, mixed-phase physicochemical processes has resulted in an advanced understanding of the evolution of atmospheric systems but has also introduced a level of complexity that few atmospheric models were designed to handle. We present a flexible treatment for multiphase chemical processes for models of diverse scale, from box up to global models. This enables users to build a customized multiphase mechanism that is accessible to a much wider community.

Paleoclimate research requires better numerical model tools to explore interactions among the cryosphere, atmosphere, ocean and land surface. To explore those interactions, this study offers a tool, the GREB-ISM, which can be run for 2 million model years within 1 month on a personal computer. A series of experiments show that the GREB-ISM is able to reproduce the modern ice sheet distribution as well as classic climate oscillation features under paleoclimate conditions.

We present the dynamical core of the MPAS-Seaice model, which uses a mesh consisting of a Voronoi tessellation with polygonal cells. Such a mesh allows variable mesh resolution in different parts of the domain and the focusing of computational resources in regions of interest. We describe the velocity solver and tracer transport schemes used and examine errors generated by the model in both idealized and realistic test cases and examine the computational efficiency of the model.

Ice sheet models are used to simulate the changes of ice sheets in future but are currently often run in coarse resolution and/or with neglecting important physics to make them affordable in terms of computational costs. We conducted a study simulating the Greenland Ice Sheet in high resolution and adequate physics to test where the ISSM ice sheet code is using most time and what could be done to improve its performance for future computer architectures that allow massive parallel computing.

Computer simulations play a fundamental role in understanding volcanic phenomena. The growing complexity of these simulations requires the development of flexible computational tools that can easily switch between sub-models and solution techniques as well as optimizations. MagmaFOAM is a newly developed library that allows for maximum flexibility for solving multiphase volcanic flows and promotes collaborative work for in-house and community model development, testing, and comparison.

The predication of PM_2.5 has been carried out using a numerical air quality model in South Korea. Despite recent progress of numerical air quality models, accurate prediction of PM_2.5 is still challenging. In this study, we developed a data-based model using a deep neural network (DNN) to overcome the limitations of numerical air quality models. The results showed that the DNN model outperformed the CMAQ when it was trained by using observation and forecasting data from the numerical models.

Full waveform inversion (FWI) is a partial-differential equation (PDE)-constrained optimization problem that is notorious for its high computational load and memory footprint. In this paper we present a method that combines recomputation with lossy compression to accelerate the computation with minimal loss of precision in the results. We show this using experiments running FWI with a variety of compression settings on a popular academic dataset.

In this study, we present a novel formulation to build a dynamical combination of models, the so-called supermodel, which needs to be trained based on data. Previously, we assumed complete and noise-free observations. Here, we move towards a realistic scenario and develop adaptations to the training methods in order to cope with sparse and noisy observations. The results are very promising and shed light on how to apply the method with state of the art general circulation models.

Mercury is one of the most toxic pollutants in the environment, and wet deposition is a major process for atmospheric mercury to enter, causing ecological and human health risks. High-mercury wet deposition in the southeastern US has been a problem for many years. Here we employed a newly developed high-resolution WRF-GC model with the capability to simulate mercury to study this problem. We conclude that deep convection caused enhanced mercury wet deposition in the southeastern US.

We present the Water Ecosystems Tool (WET) – a new modular aquatic ecosystem model configurable to a wide array of physical setups, ecosystems and research questions based on the popular FABM–PCLake model. We aim for the model to become a community staple, thus helping to consolidate the state of the art under a few flexible models, with the aim of improving comparability across studies and preventing the re-inventions of the wheel that are common to our scientific modeling community.

In systems such as atmospheric clouds, droplets undergo growth through condensation of vapor. The broadness of the resultant size spectrum of droplets influences precipitation likelihood and the radiative properties of clouds. One of the inherent limitations of simulations of the problem is the so-called numerical diffusion causing overestimation of the spectrum width, hence the term numerical broadening. In the paper, we take a closer look at one of the algorithms used in this context: MPDATA.

We describe the coupled bio-physical model ECOSMO II(CHL), which is used for regional configurations for the North Atlantic and the Arctic hind-casting and operational purposes. The model is consistent with the large-scale climatological nutrient settings and is capable of representing regional and seasonal changes, and model primary production agrees with previous measurements. For the users of this model, this paper provides the underlying science, model evaluation and its development.

This study uses a set of deep neural networks to learn a parameterization scheme from a superparameterized general circulation model (GCM). After being embedded in a realistically configurated GCM, the parameterization scheme performs stably in long-term climate simulations and reproduces reasonable climatology and climate variability. This success is the first for long-term stable climate simulations using machine learning parameterization under real geographical boundary conditions.

We document the model experiments for the future climate projection by E3SMv1.0. At the highest future emission scenario, E3SMv1.0 projects a strong surface warming with rapid changes in the atmosphere, ocean, sea ice, and land runoff. Specifically, we detect a significant polar amplification and accelerated warming linked to the unmasking of the aerosol effects. The impact of greenhouse gas forcing is examined in different climate components.

The community aerosol dynamics model MAFOR includes several advanced features: coupling with an up-to-date chemistry mechanism for volatile organic compounds, a revised Brownian coagulation kernel that takes into account the fractal geometry of soot particles, a multitude of nucleation parameterizations, size-resolved partitioning of semi-volatile inorganics, and a hybrid method for the formation of secondary organic aerosols within the framework of condensation and evaporation.