Search

Exploring planetary systems similar to our solar system can provide a means to explore a large range of possibly temperate climates on Earth-like worlds. Rather than run hundreds of simulations with different eccentricities at fixed obliquities, our variable-eccentricity approach provides a means to cover an incredibly large parameter space. Herein Jupiter's orbital radius is moved substantially inward in two different scenarios, causing a forcing on Earth's eccentricity. In one case, the eccentricity of Earth varies from 0 to 0.27 over ~7000 yr for three different fixed obliquities (0{\deg}, 23{\deg}, and 45{\deg}). In another case, the eccentricity varies from 0 to 0.53 over ~9400 yr in a single case with zero obliquity. In all cases, we find that the climate remains stable, but regional habitability changes through time in unique ways. At the same time, the moist greenhouse state is approached but only when at the highest eccentricities., Comment: 10 pages, 11 Figures. Data/scripts used to generate figures & ROCKE-3D boundary condition files can be downloaded from doi:10.5281/zenodo.8398270

The nearby exoplanet Proxima Centauri b will be a prime future target for characterization, despite questions about its retention of water. Climate models with static oceans suggest that an Earth-like Proxima b could harbor a small dayside region of surface liquid water at fairly warm temperatures despite its weak instellation. We present the first 3-dimensional climate simulations of Proxima b with a dynamic ocean. We find that an ocean-covered Proxima b could have a much broader area of surface liquid water but at much colder temperatures than previously suggested, due to ocean heat transport and depression of the freezing point by salinity. Elevated greenhouse gas concentrations do not necessarily produce more open ocean area because of possible dynamic regime transitions. For an evolutionary path leading to a highly saline present ocean, Proxima b could conceivably be an inhabited, mostly open ocean planet dominated by halophilic life. For an ocean planet in 3:2 spin-orbit resonance, a permanent tropical waterbelt exists for moderate eccentricity. Simulations of Proxima Centauri b may also be a model for the habitability of planets receiving similar instellation from slightly cooler or warmer stars, e.g., in the TRAPPIST-1, LHS 1140, GJ 273, and GJ 3293 systems., Comment: Submitted to Astrobiology; 38 pages, 12 figures, 5 tables

This technical report records and discusses the Second Workshop on Sustainable Software for Science: Practice and Experiences (WSSSPE2). The report includes a description of the alternative, experimental submission and review process, two workshop keynote presentations, a series of lightning talks, a discussion on sustainability, and five discussions from the topic areas of exploring sustainability; software development experiences; credit & incentives; reproducibility & reuse & sharing; and code testing & code review. For each topic, the report includes a list of tangible actions that were proposed and that would lead to potential change. The workshop recognized that reliance on scientific software is pervasive in all areas of world-leading research today. The workshop participants then proceeded to explore different perspectives on the concept of sustainability. Key enablers and barriers of sustainable scientific software were identified from their experiences. In addition, recommendations with new requirements such as software credit files and software prize frameworks were outlined for improving practices in sustainable software engineering. There was also broad consensus that formal training in software development or engineering was rare among the practitioners. Significant strides need to be made in building a sense of community via training in software and technical practices, on increasing their size and scope, and on better integrating them directly into graduate education programs. Finally, journals can define and publish policies to improve reproducibility, whereas reviewers can insist that authors provide sufficient information and access to data and software to allow them reproduce the results in the paper. Hence a list of criteria is compiled for journals to provide to reviewers so as to make it easier to review software submitted for publication as a "Software Paper."

Atmospheric chemistry models are a central tool to study and forecast the impact of air pollution on the environment, vegetation, and human health. However, the numerical simulation of chemical kinetics is computationally expensive due to the stiffness of the system of ordinary differential equations that describes atmospheric chemistry. Here we present an alternative approach to the computation of atmospheric chemistry based on machine learning. Our training data set is produced using the NASA Goddard Earth Observing System (GEOS) model with GEOS-Chem chemistry, run on the NASA Center for Climate Simulation (NCCS) Discover supercomputing cluster on 384 Intel Xeon Haswell cores. This model spends more than 50% of total run time on solving atmospheric chemistry. The data set contains as input features the air pollution concentrations before solving the differential equations, together with some key physical parameters such as temperature and sun intensity. As target variables we define the air pollution concentrations after solving the differential equations. Using Dask-cuDF and Dask-XGBoost on the NVIDIA RAPIDS platform on 8 Tesla V100 GPUs, we generate from this training set gradient boosted decision tree models that can reproduce the simulation of chemical kinetics. We do this on the NCCS Advanced Data Analytics Platform (ADAPT) science cloud environment. Our application takes full advantage of recent advances in Dask-XGBoost, such as multi-node and multi-GPU scaling for distributed training with large data sets. The increase in training data size enabled by this is critical to capture the full range of chemical environments encountered across the globe and all annual seasons.The boosted tree models offer good predictability and show many of the features of the full chemistry reference simulation. Further improvements can be achieved through mass balance considerations and by accounting for error correlations. We incorporate the boosted tree models into the GEOS reference model using XGBoost's C API. This enables a seamless integration of the GPU trained models into GEOS-Chem, which is written in Fortran and optimized for use in a massively parallel CPU environment. We show the benefits of this approach and discuss the potential speedup of this machine learning accelerated atmospheric chemistry model.

Sensor forward models are an important tool for interpreting remote sensing observations of geophysical phenomena. By implementing a three-dimensional framework, we can simulate and analyze observations from various sensors on disparate platforms. To demonstrate our model framework, we simulate observations from the Olympic Mountains Experiment (OLYMPEX). The use of cloud model simulations allows us to understand sensor response to cloud ice, falling snow, and other processes and features, and the application of model tools to observations allows us to quantify precipitation.MIIST 3D Forward ModelThe Multi-Instrument Inverse Solver Testbed(MIIST) uses the Atmospheric Radiative TransferSimulator (ARTS) for solving the vector radiativetransfer (RT) equation in up to three spatialdimensions within a spherical geometry• Gas absorptiono Line-by-line calculationso Fast transmittance tables• Hydrometeor scattering solverso Discrete ordinateo RT4 (Evans, 1D)o Radar Single Scattering (1D or 3D)o Monte Carlo (3D)Scattering TablesHigh-fidelity hydrometeor scatteringtables are necessary for accurateand consistent forward modeling ofmulti-frequency observations• Requires full Stokes matriceso And absorption vector• Randomly oriented particleso Discrete Dipole Approximationo Characteristic Basis Function Method(coming soon)• Horizontally-oriented plateso Invariant Imbedding T-matrix MethodCloud Resolving SimulationsCloud resolving simulations (e.g.,NU-WRF) supply output consistentwith ARTS needs• Atmospheric Informationo Temperatureo Pressure / heighto Water vapor• Hydrometeor Profileso ARTS architecture ripe for explicit binmicrophysics• Examples use Morrison 2M schemeThe Olympic Mountains Experiment (OLYMPEX)Validation for GPM of mid-latitudefrontal systems approaching nearcoastalmountains from the ocean• Large collection of ground-based andairborne sensorso Radarso Radiometerso In situ• Contemporaneous with RADEXo Two sets of radar at same frequenciesRadiometer Simulation (3 km NUWRF, 20151203, 15:00)2018.12.14 7Simulate 166 GHz polarizationdifference• Corresponds to the presence of aligned icecrystals• Look at trends for both simulations andobservations• Simulations can tolerate lower resolutiono Larger domainSimulations from Observations: OLYMPEXSimulate sensor response usinggeophysical retrievals as input• Single frequency radar retrievals• Multiple scattering enhancementapparent at W band• Spatially dependent phenomenonModeling Application: 1D Retrievals03 December 2015• DC-8 and ER-2 flightso Focus on APR-3 (DC-8)• Citationo Stacked microphysics legso Qualitative comparisonso Range of frozen habitso Presence of supercooledliquid cloudsResults• Retrievals match probeso Good qualitative match• Bands of increasedreflectivity correspond tolarge Dm and highaggregate fraction• Significant amounts ofsupercooled liquid water

In the context of high performance computing (HPC), software investments in support of text-based diagnostics, which monitor a running application, are typically limited compared to those for other types of IO. Examples of such diagnostics include reiteration of configuration parameters, progress indicators, simple metrics (e.g., mass conservation, convergence of solvers, etc.), and timers. To some degree, this difference in priority is justifiable as other forms of output are the primary products of a scientific model and, due to their large data volume, much more likely to be a significant performance concern. In contrast, text-based diagnostic content is generally not shared beyond the individual or group running an application and is most often used to troubleshoot when something goes wrong. We suggest that a more systematic approach enabled by a logging facility (or logger) similar to those routinely used by many communities would provide significant value to complex scientific applications. In the context of high-performance computing, an appropriate logger would provide specialized support for distributed and shared-memory parallelism and have low performance overhead. In this paper, we present our prototype implementation of pFlogger a parallel Fortran-based logging framework, and assess its suitability for use in a complex scientific application.

Developers working in Computational Science & Engineering (CSE)/High Performance Computing (HPC) must contend with constant change due to advances in computing technology and science. Test Driven Development (TDD) is a methodology that mitigates software development risks due to change at the cost of adding comprehensive and continuous testing to the development process. Testing frameworks tailored for CSE/HPC, like pFUnit, can lower the barriers to such testing, yet CSE software faces unique constraints foreign to the broader software engineering community. Effective testing of numerical software requires a comprehensive suite of oracles, i.e., use cases with known answers, as well as robust estimates for the unavoidable numerical errors associated with implementation with finite-precision arithmetic. At first glance these concerns often seem exceedingly challenging or even insurmountable for real-world scientific applications. However, we argue that this common perception is incorrect and driven by (1) a conflation between model validation and software verification and (2) the general tendency in the scientific community to develop relatively coarse-grained, large procedures that compound numerous algorithmic steps.We believe TDD can be applied routinely to numerical software if developers pursue fine-grained implementations that permit testing, neatly side-stepping concerns about needing nontrivial oracles as well as the accumulation of errors. We present an example of a successful, complex legacy CSE/HPC code whose development process shares some aspects with TDD, which we contrast with current and potential capabilities. A mix of our proposed methodology and framework support should enable everyday use of TDD by CSE-expert developers.

Pseudospectral (PS) methods possess a number of characteristics (e.g., efficiency, accuracy, natural boundary conditions) that are extremely desirable for dynamo models. Unfortunately, dynamo models based upon PS methods face a number of daunting challenges, which include exposing additional parallelism, leveraging hardware accelerators, exploiting hybrid parallelism, and improving the scalability of global memory transposes. Although these issues are a concern for most models, solutions for PS methods tend to require far more pervasive changes to underlying data and control structures. Further, improvements in performance in one model are difficult to transfer to other models, resulting in significant duplication of effort across the research community. We have developed an extensible software framework for pseudospectral methods called SpF that is intended to enable extreme scalability and optimal performance. Highlevel abstractions provided by SpF unburden applications of the responsibility of managing domain decomposition and load balance while reducing the changes in code required to adapt to new computing architectures. The key design concept in SpF is that each phase of the numerical calculation is partitioned into disjoint numerical kernels that can be performed entirely inprocessor. The granularity of domain decomposition provided by SpF is only constrained by the datalocality requirements of these kernels. SpF builds on top of optimized vendor libraries for common numerical operations such as transforms, matrix solvers, etc., but can also be configured to use open source alternatives for portability. SpF includes several alternative schemes for global data redistribution and is expected to serve as an ideal testbed for further research into optimal approaches for different network architectures. In this presentation, we will describe our experience in porting legacy pseudospectral models, MoSST and DYNAMO, to use SpF as well as present preliminary performance results provided by the improved scalability.

Test-Driven Development (TDD), a software development process that promises many advantages for developer productivity and software reliability, has become widely accepted among professional software engineers. As the name suggests, TDD practitioners alternate between writing short automated tests and producing code that passes those tests. Although this overly simplified description will undoubtedly sound prohibitively burdensome to many uninitiated developers, the advent of powerful unit-testing frameworks greatly reduces the effort required to produce and routinely execute suites of tests. By testimony, many developers find TDD to be addicting after only a few days of exposure, and find it unthinkable to return to previous practices.After a brief overview of the TDD process and my experience in applying the methodology for development activities at Goddard, I will delve more deeply into some of the challenges that are posed by numerical and scientific software as well as tools and implementation approaches that should address those challenges.

We present a description of the ModelE2 version of the Goddard Institute for Space Studies (GISS) General Circulation Model (GCM) and the configurations used in the simulations performed for the Coupled Model Intercomparison Project Phase 5 (CMIP5). We use six variations related to the treatment of the atmospheric composition, the calculation of aerosol indirect effects, and ocean model component. Specifically, we test the difference between atmospheric models that have noninteractive composition, where radiatively important aerosols and ozone are prescribed from precomputed decadal averages, and interactive versions where atmospheric chemistry and aerosols are calculated given decadally varying emissions. The impact of the first aerosol indirect effect on clouds is either specified using a simple tuning, or parameterized using a cloud microphysics scheme. We also use two dynamic ocean components: the Russell and HYbrid Coordinate Ocean Model (HYCOM) which differ significantly in their basic formulations and grid. Results are presented for the climatological means over the satellite era (1980-2004) taken from transient simulations starting from the preindustrial (1850) driven by estimates of appropriate forcings over the 20th Century. Differences in base climate and variability related to the choice of ocean model are large, indicating an important structural uncertainty. The impact of interactive atmospheric composition on the climatology is relatively small except in regions such as the lower stratosphere, where ozone plays an important role, and the tropics, where aerosol changes affect the hydrological cycle and cloud cover. While key improvements over previous versions of the model are evident, these are not uniform across all metrics.

Observations of climate change during the CMIP5 extended historical period (1850-2012) are compared to trends simulated by six versions of the NASA Goddard Institute for Space Studies ModelE2 Earth System Model. The six models are constructed from three versions of the ModelE2 atmospheric general circulation model, distinguished by their treatment of atmospheric composition and the aerosol indirect effect, combined with two ocean general circulation models, HYCOM and Russell. Forcings that perturb the model climate during the historical period are described. Five-member ensemble averages from each of the six versions of ModelE2 simulate trends of surface air temperature, atmospheric temperature, sea ice and ocean heat content that are in general agreement with observed trends, although simulated warming is slightly excessive within the past decade. Only simulations that include increasing concentrations of long-lived greenhouse gases match the warming observed during the twentieth century. Differences in twentieth-century warming among the six model versions can be attributed to differences in climate sensitivity, aerosol and ozone forcing, and heat uptake by the deep ocean. Coupled models with HYCOM export less heat to the deep ocean, associated with reduced surface warming in regions of deepwater formation, but greater warming elsewhere at high latitudes along with reduced sea ice. All ensembles show twentieth-century annular trends toward reduced surface pressure at southern high latitudes and a poleward shift of the midlatitude westerlies, consistent with observations.

Test-Driven Development (TDD) is a software development process that promises many advantages for developer productivity and has become widely accepted among professional software engineers. As the name suggests, TDD practitioners alternate between writing short automated tests and producing code that passes those tests. Although this overly simplified description will undoubtedly sound prohibitively burdensome to many uninitiated developers, the advent of powerful unit-testing frameworks greatly reduces the effort required to produce and routinely execute suites of tests. By testimony, many developers find TDD to be addicting after only a few days of exposure, and find it unthinkable to return to previous practices. Of course, scientific/technical software differs from other software categories in a number of important respects, but I nonetheless believe that TDD is quite applicable to the development of such software and has the potential to significantly improve programmer productivity and code quality within the scientific community. After a detailed introduction to TDD, I will present the experience within the Software Systems Support Office (SSSO) in applying the technique to various scientific applications. This discussion will emphasize the various direct and indirect benefits as well as some of the difficulties and limitations of the methodology. I will conclude with a brief description of pFUnit, a unit testing framework I co-developed to support test-driven development of parallel Fortran applications.

MapReduce is an approach to high-performance analytics that may be useful to data intensive problems in climate research. It offers an analysis paradigm that uses clusters of computers and combines distributed storage of large data sets with parallel computation. We are particularly interested in the potential of MapReduce to speed up basic operations common to a wide range of analyses. In order to evaluate this potential, we are prototyping a series of canonical MapReduce operations over a test suite of observational and climate simulation datasets. Our initial focus has been on averaging operations over arbitrary spatial and temporal extents within Modern Era Retrospective- Analysis for Research and Applications (MERRA) data. Preliminary results suggest this approach can improve efficiencies within data intensive analytic workflows.

A large portion of Earth Science investigations is phenomenon- or event-based, such as the studies of Rossby waves, mesoscale convective systems, and tropical cyclones. However, except for a few high-impact phenomena, e.g. tropical cyclones, comprehensive records are absent for the occurrences or events of these phenomena. Phenomenon-based studies therefore often focus on a few prominent cases while the lesser ones are overlooked. Without an automated means to gather the events, comprehensive investigation of a phenomenon is at least time-consuming if not impossible. An Earth Science event (ES event) is defined here as an episode of an Earth Science phenomenon. A cumulus cloud, a thunderstorm shower, a rogue wave, a tornado, an earthquake, a tsunami, a hurricane, or an EI Nino, is each an episode of a named ES phenomenon," and, from the small and insignificant to the large and potent, all are examples of ES events. An ES event has a finite duration and an associated geolocation as a function of time; its therefore an entity in four-dimensional . (4D) spatiotemporal space. The interests of Earth scientists typically rivet on Earth Science phenomena with potential to cause massive economic disruption or loss of life, but broader scientific curiosity also drives the study of phenomena that pose no immediate danger. We generally gain understanding of a given phenomenon by observing and studying individual events - usually beginning by identifying the occurrences of these events. Once representative events are identified or found, we must locate associated observed or simulated data prior to commencing analysis and concerted studies of the phenomenon. Knowledge concerning the phenomenon can accumulate only after analysis has started. However, except for a few high-impact phenomena. such as tropical cyclones and tornadoes, finding events and locating associated data currently may take a prohibitive amount of time and effort on the part of an individual investigator. And even for these high-impact phenomena, the availability of comprehensive records is still only a recent development. A major reason for the lack of comprehensive ,records for the majority of the ES phenomena is the perception that they do not pose immediate and/or severe threat to life and property and are thus not consistently tracked. monitored, and catalogued. Many phenomena even lack commonly accepted criteria for definitions. However. the lack of comprehensive records is also due to the increasingly prohibitive volume of observations and model data that must be examined. NASA Earth Observing System Data Information System (EOSDIS) alone archives several petabytes (PB) of satellite remote sensing data and steadily increases. All of these factors contribute to the difficulty of methodically identifying events corresponding to a given phenomenon and significantly impede systematic investigations. In the following we present a couple motivating scenarios, demonstrating the issues faced by Earth scientists studying ES phenomena.

Over the past 30 years most climate models have grown from relatively simple representations of a few atmospheric processes to a complex multi-disciplinary system. Computer infrastructure over that period has gone from punch card mainframes to modem parallel clusters. Model implementations have become complex, brittle, and increasingly difficult to extend and maintain. Existing verification processes for model implementations rely almost exclusively upon some combination of detailed analysis of output from full climate simulations and system-level regression tests. In additional to being quite costly in terms of developer time and computing resources, these testing methodologies are limited in terms of the types of defects that can be detected, isolated and diagnosed. Mitigating these weaknesses of coarse-grained testing with finer-grained "unit" tests has been perceived as cumbersome and counter-productive. In the commercial software sector, recent advances in tools and methodology have led to a renaissance for systematic fine-grained testing. We discuss the availability of analogous tools for scientific software and examine benefits that similar testing methodologies could bring to climate modeling software. We describe the unique challenges faced when testing complex numerical algorithms and suggest techniques to minimize and/or eliminate the difficulties.