We propose the design and implementation of an ensemble data assimilation framework for weather prediction at a high resolution and with a large ensemble size. We consider the deployment of this framework on the data throughput of file I/O and multi-node communication. With regard to high-performance computing systems, where data throughput performance increases at a slower rate than computational performance, our new framework promises drastic reduction of total execution time.

In ice-flow modelling, computing in 3-D requires a lot of resources, but 2-D models lack physical likelihood when the flow is diverging. That is why 2-D models accounting for the divergence, so-called 2.5-D models, are an interesting trade-off. However, the applicability of these 2.5-D models has never been systematically examined. We show that these models are ineffective in the case of highly diverging flows, but also for varying temperature, which was not suspected.

For many applications in geoscientific modelling hourly meteorological time series are required, which generally cover shorter periods of time compared to daily time series. We present an open-source MEteoroLOgical observation time series DISaggregation Tool (MELODIST) capable of disaggregating temperature, precipitation, humidity, wind speed, and shortwave radiation (i.e. making 24 out of 1 value). Results indicate a good reconstruction of diurnal features at five sites in different climates.

In this paper, two global climate models (NorESM1-M and CCSM4) are compared to an observational-based data set (ERA-Interim) for their ability to simulate historical Arctic storminess in autumn. With this in hand, the models are run through the 21st century. We find an overall significant increase in precipitation expected, with generally fewer and weaker storms in midlatitudes and partly more and stronger storms in high-latitudes. The tendencies are strongest in areas of Arctic sea ice retreat.

This paper uses observed features of the global carbon cycle to constrain how much carbon the land should take up in an Earth system model in response to increasing fossil fuel CO₂ emissions since the start of the industrial era. These models are the only tool available to us for projecting future climate change. Despite their uncertainties, if current observations can be used to constrain models then more confidence can be places in models' future climate change projections.

The MCARS code creates sensor radiances using model-generated atmospheric columns and actual sensor and solar geometry. MCARS output looks like real data, so it is usable by any code that reads MODIS data. MCARS output can be used to test remote-sensing retrieval algorithms. Users know what went into creating the radiance: atmosphere, surface, clouds, and aerosols. Models can use MCARS output to create new parameterizations of relations of atmospheric physical quantities and measured radiances.

Software quality assurance is critical to detecting errors in large, complex climate simulation codes. We focus on ocean model simulation data in the context of an ensemble-based statistical consistency testing approach developed for atmospheric data. Because ocean and atmosphere models have differing characteristics, we develop a new statistical tool to evaluate ocean model simulation data that provide a simple, subjective, and systematic way to detect errors and instil model confidence.

We have developed a new metric for climate model evaluation that quantifies the significance of modeling errors across multiple fields. Our approach dramatically reduces the amount of data that is required to evaluate field and space dependencies and increases the community's potential to make use of the extremely valuable but limited satellite observational record. Our objective is to improve the strategies that currently exist for more formal data-driven model development.

Dynamic global vegetation models (DGVMs) are used to predict the response of vegetation to climate change. We improved the representation of carbon uptake by ecosystems in a DGVM by including a wider range of trade-offs between nutrient allocation to photosynthetic capacity and leaf structure, based on observed plant traits from a worldwide data base. The improved model has higher rates of photosynthesis and net C uptake by plants, and more closely matches observations at site and global scales.

A rigorous methodology is presented to assess numerical integration schemes for stochastic models in atmospheric dispersion known as Lagrangian particle dispersion models. A series of one-dimensional test problems modelling dispersion in the atmospheric boundary layer is used to evaluate commonly used stochastic integration schemes. The results allow for optimal time-step selection for each scheme and recommendations to be made for use in operational models.

We describe an updated scheme for prescribing stratospheric aerosol in the Community Earth System Model (CESM1). The inadequate response of the CESM1 to large volcanic disturbances to the stratospheric aerosol layer (such as the 1991 Pinatubo eruption) in comparison to observations motivates the need for a new parameterization. Simulations utilizing the new scheme successfully reproduce the observed global mean and local stratospheric temperature response to the Pinatubo eruption.

Coupled ice sheet–ocean models capable of simulating moving grounding lines are just becoming available. Such models have a broad range of potential applications in studying the dynamics of ice sheets and glaciers, including assessing their contributions to sea level change. Here we describe the idealized experiments that make up three interrelated Model Intercomparison Projects (MIPs) for marine ice sheet models and regional ocean circulation models incorporating ice shelf cavities.

Surface water and energy fluxes are essential components of the Earth system. The present paper introduces a new framework for the estimation of surface energy and water fluxes at the land surface, which allows for temporally and spatially high resolved flux estimates at the global scale. The framework maximizes the usage of existing long-term satellite data records. Overall the results indicate very good agreement with in situ observations when compared against 49 FLUXNET stations worldwide.

The paper presents a new cloud scheme for regional climate model RegCM4.5. The new scheme treats microphysical processes occurring within stratiform clouds and with respect to the pre-existing scheme is able to allow a more physically realistic representation of cloud microphysics and distribution, improving the representation of the longwave and shortwave components of the cloud radiative forcing.

Model projections of ice sheet contribution to 21st century sea level rise are greatly affected by initial conditions. Solutions have been developed to infer the friction of the ice on its bedrock using observed surface velocities. A drawback of these methods is that remaining uncertainties, especially in the bedrock elevation, lead to non-physical ice flux divergence anomalies. Here, we compare two different solutions able to infer both bedrock friction and elevation with good performance.

This manuscript presents the experiment design for the PMIP4 Last Deglaciation Core experiment: a transient simulation of the last deglaciation, 21–9 ka. Specified model boundary conditions include time-varying orbital parameters, greenhouse gases, ice sheets, ice meltwater fluxes and other geographical changes (provided for 26–0 ka). The context of the experiment and the choices for the boundary conditions are explained, along with the future direction of the working group.